Image browsing and natural language paraphrases of semantic web annotations

合集下载

2007-5-“Image” metaphors and connotations in everyday language

Annual Review of Cognitive Linguistics 5 (2007), 173–192. issn ‒ / e-issn – © John Benjamins Publishing Company
174 Alice Deignan
(2) “But as the federal powers dug in their heels against change and violence increased…”
“Image” metaphors and connotations in everyday language
Alice Deignan
University of Leeds, UK
In this paper, I argue that the general notion of an image metaphor, which has been traditionally confined to so-called “one-shot metaphors”, as used in literary and poetic language, could be expanded to describe many expressions that are found in everyday language. Following Caballero (2003a), I argue that the division in cognitive linguistics of metaphors into “image” and “conceptual” is over-simplistic. I show that many of the most frequent metaphors in my data have characteristics which would qualify them for inclusion in both categories. I also argue that connotational meaning is an important characteristic of these expressions, unifying their literal and non-literal meanings. A detailed analysis of the Bank of English corpus concordance for heel shows the numerical importance of such metaphors. I refer to research into metaphor that takes an emergentist perspective, and which has led a number of other existing distinctions to be questioned. I argue that these expressions, termed “metaphoremes”, which are difficult to classify using existing distinctions, should be regarded as prototypical on the grounds of their frequency, rather than as anomalous. Keywords: metaphor, image, metonymy, emergentist, collocation, conceptual, corpus, concordance

基于无序图像的三维建模方法

Int J Comput VisDOI10.1007/s11263-007-0107-3Modeling the World from Internet Photo Collections Noah Snavely·Steven M.Seitz·Richard SzeliskiReceived:30January2007/Accepted:31October2007©Springer Science+Business Media,LLC2007Abstract There are billions of photographs on the Inter-net,comprising the largest and most diverse photo collec-tion ever assembled.How can computer vision researchers exploit this imagery?This paper explores this question from the standpoint of3D scene modeling and visualization.We present structure-from-motion and image-based rendering algorithms that operate on hundreds of images downloaded as a result of keyword-based image search queries like “Notre Dame”or“Trevi Fountain.”This approach,which we call Photo Tourism,has enabled reconstructions of nu-merous well-known world sites.This paper presents these algorithms and results as aﬁrst step towards3D modeling of the world’s well-photographed sites,cities,and landscapes from Internet imagery,and discusses key open problems and challenges for the research community.Keywords Structure from motion·3D scene analysis·Internet imagery·Photo browsers·3D navigation1IntroductionMost of the world’s signiﬁcant sites have been photographed under many different conditions,both from the ground and from the air.For example,a Google image search for“Notre Dame”returns over one million hits(as of September, 2007),showing the cathedral from almost every conceivable viewing position and angle,different times of day and night, N.Snavely( )·S.M.SeitzUniversity of Washington,Seattle,WA,USAe-mail:snavely@R.SzeliskiMicrosoft Research,Redmond,WA,USA and changes in season,weather,and decade.Furthermore, entire cities are now being captured at street level and from a birds-eye perspective(e.g.,Windows Live Local,1,2and Google Streetview3),and from satellite or aerial views(e.g., Google4).The availability of such rich imagery of large parts of the earth’s surface under many different viewing conditions presents enormous opportunities,both in computer vision research and for practical applications.From the standpoint of shape modeling research,Internet imagery presents the ultimate data set,which should enable modeling a signiﬁ-cant portion of the world’s surface geometry at high resolu-tion.As the largest,most diverse set of images ever assem-bled,Internet imagery provides deep insights into the space of natural images and a rich source of statistics and priors for modeling scene appearance.Furthermore,Internet imagery provides an ideal test bed for developing robust and gen-eral computer vision algorithms that can work effectively “in the wild.”In turn,algorithms that operate effectively on such imagery will enable a host of important applications, ranging from3D visualization,localization,communication (media sharing),and recognition,that go well beyond tradi-tional computer vision problems and can have broad impacts for the population at large.To date,this imagery is almost completely untapped and unexploited by computer vision researchers.A major rea-son is that the imagery is not in a form that is amenable to processing,at least by traditional methods:the images are 1Windows Live Local,.2Windows Live Local—Virtual Earth Technology Preview,http:// .3Google Maps,.4Google Maps,.Int J Comput Visunorganized,uncalibrated,with widely variable and uncon-trolled illumination,resolution,and image quality.Develop-ing computer vision techniques that can operate effectively with such imagery has been a major challenge for the re-search community.Within this scope,one key challenge is registration,i.e.,ﬁguring out correspondences between im-ages,and how they relate to one another in a common3D coordinate system(structure from motion).While a lot of progress has been made in these areas in the last two decades (Sect.2),many challenging open problems remain.In this paper we focus on the problem of geometrically registering Internet imagery and a number of applications that this enables.As such,weﬁrst review the state of the art and then present someﬁrst steps towards solving this problem along with a visualization front-end that we call Photo Tourism(Snavely et al.2006).We then present a set of open research problems for theﬁeld,including the cre-ation of more efﬁcient correspondence and reconstruction techniques for extremely large image data sets.This paper expands on the work originally presented in(Snavely et al. 2006)with many new reconstructions and visualizations of algorithm behavior across datasets,as well as a brief dis-cussion of Photosynth,a Technology Preview by Microsoft Live Labs,based largely on(Snavely et al.2006).We also present a more complete related work section and add a broad discussion of open research challenges for theﬁeld. Videos of our system,along with additional supplementary material,can be found on our Photo Tourism project Web site,.2Previous WorkThe last two decades have seen a dramatic increase in the capabilities of3D computer vision algorithms.These in-clude advances in feature correspondence,structure from motion,and image-based modeling.Concurrently,image-based rendering techniques have been developed in the com-puter graphics community,and image browsing techniques have been developed for multimedia applications.2.1Feature CorrespondenceTwenty years ago,the foundations of modern feature detec-tion and matching techniques were being laid.Lucas and Kanade(1981)had developed a patch tracker based on two-dimensional image statistics,while Moravec(1983)intro-duced the concept of“corner-like”feature points.Först-ner(1986)and then Harris and Stephens(1988)both pro-posedﬁnding keypoints using measures based on eigenval-ues of smoothed outer products of gradients,which are still widely used today.While these early techniques detected keypoints at a single scale,modern techniques use a quasi-continuous sampling of scale space to detect points invari-ant to changes in scale and orientation(Lowe2004;Mikola-jczyk and Schmid2004)and somewhat invariant to afﬁne transformations(Baumberg2000;Kadir and Brady2001; Schaffalitzky and Zisserman2002;Mikolajczyk et al.2005).Unfortunately,early techniques relied on matching patches around the detected keypoints,which limited their range of applicability to scenes seen from similar view-points,e.g.,for aerial photogrammetry applications(Hannah 1988).If features are being tracked from frame to frame,an afﬁne extension of the basic Lucas-Kanade tracker has been shown to perform well(Shi and Tomasi1994).However,for true wide baseline matching,i.e.,the automatic matching of images taken from widely different views(Baumberg2000; Schaffalitzky and Zisserman2002;Strecha et al.2003; Tuytelaars and Van Gool2004;Matas et al.2004),(weakly) afﬁne-invariant feature descriptors must be used.Mikolajczyk et al.(2005)review some recently devel-oped view-invariant local image descriptors and experimen-tally compare their performance.In our own Photo Tourism research,we have been using Lowe’s Scale Invariant Fea-ture Transform(SIFT)(Lowe2004),which is widely used by others and is known to perform well over a reasonable range of viewpoint variation.2.2Structure from MotionThe late1980s also saw the development of effective struc-ture from motion techniques,which aim to simultaneously reconstruct the unknown3D scene structure and camera positions and orientations from a set of feature correspon-dences.While Longuet-Higgins(1981)introduced a still widely used two-frame relative orientation technique in 1981,the development of multi-frame structure from mo-tion techniques,including factorization methods(Tomasi and Kanade1992)and global optimization techniques(Spet-sakis and Aloimonos1991;Szeliski and Kang1994;Olien-sis1999)occurred quite a bit later.More recently,related techniques from photogrammetry such as bundle adjustment(Triggs et al.1999)(with related sparse matrix techniques,Szeliski and Kang1994)have made their way into computer vision and are now regarded as the gold standard for performing optimal3D reconstruc-tion from correspondences(Hartley and Zisserman2004).For situations where the camera calibration parameters are unknown,self-calibration techniques,whichﬁrst esti-mate a projective reconstruction of the3D world and then perform a metric upgrade have proven to be successful (Pollefeys et al.1999;Pollefeys and Van Gool2002).In our own work(Sect.4.2),we have found that the simpler approach of simply estimating each camera’s focal length as part of the bundle adjustment process seems to produce good results.Int J Comput VisThe SfM approach used in this paper is similar to that of Brown and Lowe(2005),with several modiﬁcations to improve robustness over a variety of data sets.These in-clude initializing new cameras using pose estimation,to help avoid local minima;a different heuristic for selecting the initial two images for SfM;checking that reconstructed points are well-conditioned before adding them to the scene; and using focal length information from image EXIF tags. Schaffalitzky and Zisserman(2002)present another related technique for reconstructing unordered image sets,concen-trating on efﬁciently matching interest points between im-ages.Vergauwen and Van Gool have developed a similar approach(Vergauwen and Van Gool2006)and are hosting a web-based reconstruction service for use in cultural heritage applications5.Fitzgibbon and Zisserman(1998)and Nistér (2000)prefer a bottom-up approach,where small subsets of images are matched to each other and then merged in an agglomerative fashion into a complete3D reconstruction. While all of these approaches address the same SfM prob-lem that we do,they were tested on much simpler datasets with more limited variation in imaging conditions.Our pa-per marks theﬁrst successful demonstration of SfM tech-niques applied to the kinds of real-world image sets found on Google and Flickr.For instance,our typical image set has photos from hundreds of different cameras,zoom levels, resolutions,different times of day or seasons,illumination, weather,and differing amounts of occlusion.2.3Image-Based ModelingIn recent years,computer vision techniques such as structure from motion and model-based reconstruction have gained traction in the computer graphicsﬁeld under the name of image-based modeling.IBM is the process of creating three-dimensional models from a collection of input images(De-bevec et al.1996;Grzeszczuk2002;Pollefeys et al.2004).One particular application of IBM has been the cre-ation of large scale architectural models.Notable exam-ples include the semi-automatic Façade system(Debevec et al.1996),which was used to reconstruct compellingﬂy-throughs of the University of California Berkeley campus; automatic architecture reconstruction systems such as that of Dick et al.(2004);and the MIT City Scanning Project (Teller et al.2003),which captured thousands of calibrated images from an instrumented rig to construct a3D model of the MIT campus.There are also several ongoing academic and commercial projects focused on large-scale urban scene reconstruction.These efforts include the4D Cities project (Schindler et al.2007),which aims to create a spatial-temporal model of Atlanta from historical photographs;the 5Epoch3D Webservice,http://homes.esat.kuleuven.be/~visit3d/ webservice/html/.Stanford CityBlock Project(Román et al.2004),which uses video of city blocks to create multi-perspective strip images; and the UrbanScape project of Akbarzadeh et al.(2006). Our work differs from these previous approaches in that we only reconstruct a sparse3D model of the world,since our emphasis is more on creating smooth3D transitions be-tween photographs rather than interactively visualizing a3D world.2.4Image-Based RenderingTheﬁeld of image-based rendering(IBR)is devoted to the problem of synthesizing new views of a scene from a set of input photographs.A forerunner to thisﬁeld was the groundbreaking Aspen MovieMap project(Lippman1980), in which thousands of images of Aspen Colorado were cap-tured from a moving car,registered to a street map of the city,and stored on laserdisc.A user interface enabled in-teractively moving through the images as a function of the desired path of the user.Additional features included a navi-gation map of the city overlaid on the image display,and the ability to touch any building in the currentﬁeld of view and jump to a facade of that building.The system also allowed attaching metadata such as restaurant menus and historical images with individual buildings.Recently,several compa-nies,such as Google6and EveryScape7have begun creating similar“surrogate travel”applications that can be viewed in a web browser.Our work can be seen as a way to automati-cally create MovieMaps from unorganized collections of im-ages.(In contrast,the Aspen MovieMap involved a team of over a dozen people working over a few years.)A number of our visualization,navigation,and annotation capabilities are similar to those in the original MovieMap work,but in an improved and generalized form.More recent work in IBR has focused on techniques for new view synthesis,e.g.,(Chen and Williams1993; McMillan and Bishop1995;Gortler et al.1996;Levoy and Hanrahan1996;Seitz and Dyer1996;Aliaga et al.2003; Zitnick et al.2004;Buehler et al.2001).In terms of appli-cations,Aliaga et al.’s(2003)Sea of Images work is perhaps closest to ours in its use of a large collection of images taken throughout an architectural space;the same authors address the problem of computing consistent feature matches across multiple images for the purposes of IBR(Aliaga et al.2003). However,our images are casually acquired by different pho-tographers,rather than being taken on aﬁxed grid with a guided robot.In contrast to most prior work in IBR,our objective is not to synthesize a photo-realistic view of the world from all viewpoints per se,but to browse a speciﬁc collection of 6Google Maps,.7Everyscape,.Int J Comput Visphotographs in a3D spatial context that gives a sense of the geometry of the underlying scene.Our approach there-fore uses an approximate plane-based view interpolation method and a non-photorealistic rendering of background scene structures.As such,we side-step the more challenging problems of reconstructing full surface models(Debevec et al.1996;Teller et al.2003),lightﬁelds(Gortler et al.1996; Levoy and Hanrahan1996),or pixel-accurate view inter-polations(Chen and Williams1993;McMillan and Bishop 1995;Seitz and Dyer1996;Zitnick et al.2004).The bene-ﬁt of doing this is that we are able to operate robustly with input imagery that is beyond the scope of previous IBM and IBR techniques.2.5Image Browsing,Retrieval,and AnnotationThere are many techniques and commercial products for browsing sets of photos and much research on the subject of how people tend to organize photos,e.g.,(Rodden and Wood2003).Many of these techniques use metadata,such as keywords,photographer,or time,as a basis of photo or-ganization(Cooper et al.2003).There has recently been growing interest in using geo-location information to facilitate photo browsing.In particu-lar,the World-Wide Media Exchange(WWMX)(Toyama et al.2003)arranges images on an interactive2D map.Photo-Compas(Naaman et al.2004)clusters images based on time and location.Realityﬂythrough(McCurdy and Griswold 2005)uses interface ideas similar to ours for exploring video from camcorders instrumented with GPS and tilt sensors, and Kadobayashi and Tanaka(2005)present an interface for retrieving images using proximity to a virtual camera.In Photowalker(Tanaka et al.2002),a user can manually au-thor a walkthrough of a scene by specifying transitions be-tween pairs of images in a collection.In these systems,loca-tion is obtained from GPS or is manually speciﬁed.Because our approach does not require GPS or other instrumentation, it has the advantage of being applicable to existing image databases and photographs from the Internet.Furthermore, many of the navigation features of our approach exploit the computation of image feature correspondences and sparse 3D geometry,and therefore go beyond what has been possi-ble in these previous location-based systems.Many techniques also exist for the related task of retriev-ing images from a database.One particular system related to our work is Video Google(Sivic and Zisserman2003)(not to be confused with Google’s own video search),which al-lows a user to select a query object in one frame of video and efﬁcientlyﬁnd that object in other frames.Our object-based navigation mode uses a similar idea,but extended to the3D domain.A number of researchers have studied techniques for au-tomatic and semi-automatic image annotation,and annota-tion transfer in particular.The LOCALE system(Naaman et al.2003)uses proximity to transfer labels between geo-referenced photographs.An advantage of the annotation ca-pabilities of our system is that our feature correspondences enable transfer at muchﬁner granularity;we can transfer annotations of speciﬁc objects and regions between images, taking into account occlusions and the motions of these ob-jects under changes in viewpoint.This goal is similar to that of augmented reality(AR)approaches(e.g.,Feiner et al. 1997),which also seek to annotate images.While most AR methods register a3D computer-generated model to an im-age,we instead transfer2D image annotations to other im-ages.Generating annotation content is therefore much eas-ier.(We can,in fact,import existing annotations from pop-ular services like Flickr.)Annotation transfer has been also explored for video sequences(Irani and Anandan1998).Finally,Johansson and Cipolla(2002)have developed a system where a user can take a photograph,upload it to a server where it is compared to an image database,and re-ceive location information.Our system also supports this application in addition to many other capabilities(visual-ization,navigation,annotation,etc.).3OverviewOur objective is to geometrically register large photo col-lections from the Internet and other sources,and to use the resulting3D camera and scene information to facili-tate a number of applications in visualization,localization, image browsing,and other areas.This section provides an overview of our approach and summarizes the rest of the paper.The primary technical challenge is to robustly match and reconstruct3D information from hundreds or thousands of images that exhibit large variations in viewpoint,illumina-tion,weather conditions,resolution,etc.,and may contain signiﬁcant clutter and outliers.This kind of variation is what makes Internet imagery(i.e.,images returned by Internet image search queries from sites such as Flickr and Google) so challenging to work with.In tackling this problem,we take advantage of two recent breakthroughs in computer vision,namely feature-matching and structure from motion,as reviewed in Sect.2.The back-bone of our work is a robust SfM approach that reconstructs 3D camera positions and sparse point geometry for large datasets and has yielded reconstructions for dozens of fa-mous sites ranging from Notre Dame Cathedral to the Great Wall of China.Section4describes this approach in detail, as well as methods for aligning reconstructions to satellite and map data to obtain geo-referenced camera positions and geometry.One of the most exciting applications for these recon-structions is3D scene visualization.However,the sparseInt J Comput Vispoints produced by SfM methods are by themselves very limited and do not directly produce compelling scene ren-derings.Nevertheless,we demonstrate that this sparse SfM-derived geometry and camera information,along with mor-phing and non-photorealistic rendering techniques,is sufﬁ-cient to provide compelling view interpolations as described in5.Leveraging this capability,Section6describes a novel photo explorer interface for browsing large collections of photographs in which the user can virtually explore the3D space by moving from one image to another.Often,we are interested in learning more about the con-tent of an image,e.g.,“which statue is this?”or“when was this building constructed?”A great deal of annotated image content of this form already exists in guidebooks,maps,and Internet resources such as Wikipedia8and Flickr.However, the image you may be viewing at any particular time(e.g., from your cell phone camera)may not have such annota-tions.A key feature of our system is the ability to transfer annotations automatically between images,so that informa-tion about an object in one image is propagated to all other images that contain the same object(Sect.7).Section8presents extensive results on11scenes,with visualizations and an analysis of the matching and recon-struction results for these scenes.We also brieﬂy describe Photosynth,a related3D image browsing tool developed by Microsoft Live Labs that is based on techniques from this paper,but also adds a number of interesting new elements. Finally,we conclude with a set of research challenges for the community in Sect.9.4Reconstructing Cameras and Sparse GeometryThe visualization and browsing components of our system require accurate information about the relative location,ori-entation,and intrinsic parameters such as focal lengths for each photograph in a collection,as well as sparse3D scene geometry.A few features of our system require the absolute locations of the cameras,in a geo-referenced coordinate frame.Some of this information can be provided with GPS devices and electronic compasses,but the vast majority of existing photographs lack such information.Many digital cameras embed focal length and other information in the EXIF tags of imageﬁles.These values are useful for ini-tialization,but are sometimes inaccurate.In our system,we do not rely on the camera or any other piece of equipment to provide us with location,orientation, or geometry.Instead,we compute this information from the images themselves using computer vision techniques.We ﬁrst detect feature points in each image,then match feature points between pairs of images,andﬁnally run an iterative, 8Wikipedia,.robust SfM procedure to recover the camera parameters.Be-cause SfM only estimates the relative position of each cam-era,and we are also interested in absolute coordinates(e.g., latitude and longitude),we use an interactive technique to register the recovered cameras to an overhead map.Each of these steps is described in the following subsections.4.1Keypoint Detection and MatchingTheﬁrst step is toﬁnd feature points in each image.We use the SIFT keypoint detector(Lowe2004),because of its good invariance to image transformations.Other feature de-tectors could also potentially be used;several detectors are compared in the work of Mikolajczyk et al.(2005).In addi-tion to the keypoint locations themselves,SIFT provides a local descriptor for each keypoint.A typical image contains several thousand SIFT keypoints.Next,for each pair of images,we match keypoint descrip-tors between the pair,using the approximate nearest neigh-bors(ANN)kd-tree package of Arya et al.(1998).To match keypoints between two images I and J,we create a kd-tree from the feature descriptors in J,then,for each feature in I weﬁnd the nearest neighbor in J using the kd-tree.For efﬁciency,we use ANN’s priority search algorithm,limiting each query to visit a maximum of200bins in the tree.Rather than classifying false matches by thresholding the distance to the nearest neighbor,we use the ratio test described by Lowe(2004):for a feature descriptor in I,weﬁnd the two nearest neighbors in J,with distances d1and d2,then accept the match if d1d2<0.6.If more than one feature in I matches the same feature in J,we remove all of these matches,as some of them must be spurious.After matching features for an image pair(I,J),we robustly estimate a fundamental matrix for the pair us-ing RANSAC(Fischler and Bolles1981).During each RANSAC iteration,we compute a candidate fundamental matrix using the eight-point algorithm(Hartley and Zis-serman2004),normalizing the problem to improve robust-ness to noise(Hartley1997).We set the RANSAC outlier threshold to be0.6%of the maximum image dimension,i.e., 0.006max(image width,image height)(about six pixels for a1024×768image).The F-matrix returned by RANSAC is reﬁned by running the Levenberg-Marquardt algorithm(No-cedal and Wright1999)on the eight parameters of the F-matrix,minimizing errors for all the inliers to the F-matrix. Finally,we remove matches that are outliers to the recov-ered F-matrix using the above threshold.If the number of remaining matches is less than twenty,we remove all of the matches from consideration.Afterﬁnding a set of geometrically consistent matches between each image pair,we organize the matches into tracks,where a track is a connected set of matching key-points across multiple images.If a track contains more thanInt J Comput VisFig.1Photo connectivity graph.This graph contains a node for each image in a set of photos of the Trevi Fountain, with an edge between each pair of photos with matching features.The size of a node is proportional to its degree.There are two dominant clusters corresponding to day(a)and night time(d)photos.Similar views of the facade cluster together in the center,while nodes in the periphery,e.g.,(b) and(c),are more unusual(often close-up)viewsone keypoint in the same image,it is deemed inconsistent. We keep consistent tracks containing at least two keypoints for the next phase of the reconstruction procedure.Once correspondences are found,we can construct an im-age connectivity graph,in which each image is a node and an edge exists between any pair of images with matching features.A visualization of an example connectivity graph for the Trevi Fountain is Fig.1.This graph embedding was created with the neato tool in the Graphviz toolkit.9Neato represents the graph as a mass-spring system and solves for an embedding whose energy is a local minimum.The image connectivity graph of this photo set has sev-eral distinct features.The large,dense cluster in the cen-ter of the graph consists of photos that are all fairly wide-angle,frontal,well-lit shots of the fountain(e.g.,image(a)). Other images,including the“leaf”nodes(e.g.,images(b) and(c))and night time images(e.g.,image(d)),are more loosely connected to this core set.Other connectivity graphs are shown in Figs.9and10.4.2Structure from MotionNext,we recover a set of camera parameters(e.g.,rotation, translation,and focal length)for each image and a3D lo-cation for each track.The recovered parameters should be consistent,in that the reprojection error,i.e.,the sum of dis-tances between the projections of each track and its corre-sponding image features,is minimized.This minimization problem can formulated as a non-linear least squares prob-lem(see Appendix1)and solved using bundle adjustment. Algorithms for solving this non-linear problem,such as No-cedal and Wright(1999),are only guaranteed toﬁnd lo-cal minima,and large-scale SfM problems are particularly prone to getting stuck in bad local minima,so it is important 9Graphviz—graph visualization software,/.to provide good initial estimates of the parameters.Rather than estimating the parameters for all cameras and tracks at once,we take an incremental approach,adding in one cam-era at a time.We begin by estimating the parameters of a single pair of cameras.This initial pair should have a large number of matches,but also have a large baseline,so that the ini-tial two-frame reconstruction can be robustly estimated.We therefore choose the pair of images that has the largest num-ber of matches,subject to the condition that those matches cannot be well-modeled by a single homography,to avoid degenerate cases such as coincident cameras.In particular, weﬁnd a homography between each pair of matching im-ages using RANSAC with an outlier threshold of0.4%of max(image width,image height),and store the percentage of feature matches that are inliers to the estimated homogra-phy.We select the initial image pair as that with the lowest percentage of inliers to the recovered homography,but with at least100matches.The camera parameters for this pair are estimated using Nistér’s implementation of theﬁve point al-gorithm(Nistér2004),10then the tracks visible in the two images are triangulated.Finally,we do a two frame bundle adjustment starting from this initialization.Next,we add another camera to the optimization.We select the camera that observes the largest number of tracks whose3D locations have already been estimated, and initialize the new camera’s extrinsic parameters using the direct linear transform(DLT)technique(Hartley and Zisserman2004)inside a RANSAC procedure.For this RANSAC step,we use an outlier threshold of0.4%of max(image width,image height).In addition to providing an estimate of the camera rotation and translation,the DLT technique returns an upper-triangular matrix K which can 10We only choose the initial pair among pairs for which a focal length estimate is available for both cameras,and therefore a calibrated rela-tive pose algorithm can be used.。

通信行业英语中英对照手册(i)

通信行业英语中英对照手册(I)-OCU ISDN-Office Channel Unit ISDN的局内信道单元I/O Input / Output 输入/输出I2T Intelligent Interface Technology 智能接口技术IA Information Access 信息存取IA Intelligence Appliance 智能家电IA Internal Authentication 内部认证IA Internet Address 因特网地址IAB Internet Activities Board 因特网活动委员会IAB Internet Architecture Board 因特网体系委员会IAC Image Attenuation Coefficient 图像衰减系数IAC ISDN Access Control ISDN接入控制IAD Integrated Access Device 集成接入设备IAF Image Analysis Facility 图像分析设备IAM Initial Address Management 初始地址管理IAM Initial Address Messege 初始地址消息IAN Integrated Analog Network 综合模拟网IAN Irregularly Activated Network 不规则激活网络IANA Internet Assigned Number Authority 因特网分址机构IAP Internet Access Point 网际访问点IAP Internet Access Provider 因特网接入服务供应商IAR Intelligent Automatic Rerouting 智能型自动重选路由IAS Integrated Access Server 综合接入服务器IAS Interactive Application Server 交互应用服务器IB-DCA Interence-Based Dynamic Channel Allocation 基于干扰的信道动态分配IBC Information Bearer Channel 信息承载信道IBC Integrated Broadband Communication 综合宽带通信IBCN Integrated Broadband Communication Network 综合宽带通信网IBCN International Broadband Communication Network 国际宽带通信网络IBDN Integrated Building Distribution Network 楼宇综合布线网络IBGP Internal Border Gateway Protocol 内部边界网关协议IBI Intergrated Building Intelligent 综合大楼智能化IBMS Intelligent Building Management System 智能大厦管理系统IBN Integrated Broadband Network 综合宽带网IBS Intelligent Building System 智能大厦系统IBT Internet Browsing Terminal 因特网浏览终端IBWN Indoor Broadband Wireless Network 室内宽带无线网络IC Image Check 图像检验IC Image Compression 图像压缩IC Integrated Circuit 集成电路IC Interlock Code 互锁码ICB Incoming Call Barred 来话加锁ICBI Inter-Channel inter-Block Interference 信道间信息组间的干扰ICC Instantaneous Channel Characteristics 信道瞬态特性ICC Internet Call Center 因特网呼叫中心ICDN Integrated Communication Data Network 综合通信数据网络ICE In-Circuit Emulation 在线仿真ICE InterConnect Equipment 互连设备ICE Interface Configuration Environment 接口配置环境ICH Incoming CHannel 来话信道ICI Intelligent Communications Interface 智能通信接口ICI Inter-Carrier Interference 载波间干扰ICI Inter-Channel Interference 信道间干扰ICM Image Compression Manager 图像压缩管理器ICM Incoming Call Management 来话呼叫管理ICMP Inernal Control Message Protocol 内部控制信息协议ICMP Internet Control Message Protocol 因特网控制报文协议ICP Incoming Call Packet 呼入分组信息ICP Internal Connection Protocol 内部连接协议ICP Internet Content Provider 因特网内容服务供应商ICP Interworking Control Protocol 互通控制协议ICR Initial Cell Rate 初始信元率ICS ISDN Control Sublayer ISDN控制子层ICSA International Computer Security Associatiion 国际计算机安全协会ICT InComing Trunk 来话中继ICT Information and Communication Technology 信息和通信技术ICUG International Closed User Group 国际闭合用户群ICW Internet Call Waiting 因特网呼叫等待IDA Integrated Digital Access 综合数字接入IDA Interchange of Data between Administrations 机构间的数据交换IDA Internet Direct Access 因特网直接接入IDA Intrusion Detection Agent 入侵检测代理IDARA Improved Distributed Adaptive Routing Algorithm 改进的分布式自适应路由算法IDC Internet Data Center 因特网数据中心IDCC Integrated Data Communication Channel 综合数据通信信道IDCT Inverse Discrete Consine Transform 离散余弦逆变换IDD International Direct Dialing 国际直拨IDI Initial Domain Identifier 初始域标识符IDL Interactive Distance Learning 交互式远程学习IDL Interface Definition Language 接口定义语言IDL International Data Line 国际数据线路IDLC Integrated Digital Loop Carrier 综合数字环路载波IDMS Integrated Database Management System 综合数据库管理系统IDN Integrated Data Network 综合数据网络IDN Integrated Digital Network 综合数字网IDN Intelligent Data Network 智能数据网络IDN Interactive Data Network 交互式数据网络IDN International Directory Network 国际目录网络IDNET IDentification NETwork 认证网IDP Internet Directory Provider 因特网目录服务供应商IDPR Inter-Domain Policy Routing 域间策略路由选择IDR Intermediate Data Rate 中等数据速率IDRP InterDomain Routing Protocol 域间路由选择协议IDS Intrusion Detection System 入侵检测系统IDS Isochronous Data Services 等时数据业务IDSE International Data Switching Exchange 国际数据交换机(局)IDSE Internetworking Data Switching Exchange 互连网数据交换机(局)IDSL ISDN DSL ISDN数字用户线IDSP Intelligent Dynamic Service Provisioning 智能型动态业务提供IDSS Intelligent Decision Support System 智能决策支持系统IDT Integrated Digital Terminal 综合数字终端IDT Intelligent Data Terminal 智能数据终端IDT Interactive Data Terminal 交互数据终端IDTC International Digital Transmission Center 国际数字传输中心IDU InDoor Unit 室内单元IDU Interface Data Unit 接口数据单元IEC Integrated Ethernet Chip 集成以太网电路芯片IEC InterExchange Carrier 局间载波IEC International Electrotechnical Commission 国际电工委员会IEE Institute of Electrical Engineers 电气工程师学会(英国)IEEE Institute of Electrical and Electronics Engineers 电气和电子工程师学会(美国) IEN Internet Experiment Note 因特网实验备忘录IEP Internet Equipment Provider 因特网设备供应商IEPG Internet Engineering and Planning Group 因特网工程和规划组IES ISDN Earth Station 综合业务数字网络地球站IESG Internet Engineering Steering Group 因特网工程指导组IETF Internet Engineering Task Force 因特网工程任务组IEW Intelligent and Electronic Warfare 智能和电子战IF Intermediate Frequency 中频IFD InterFace Device 接口设备IFH Intelligent Frequency Hopping 智能跳频IFIP International Federation for Information Processing 国际信息处理联合会IFITL Integrated Fiber In The Loop 综合光纤环路IFPH Inter-network FreePHone 网间被叫付费电话IFS Interactive Financial Services 交互式金融服务IFS InterFace Specification 接口规范IFS International Freephone Service 国际免费电话(被叫付费电话)IFU Interworking Functional Unit 互通功能单元IG Interactive Graphics 交互式图形IG International Gateway 国际网关IGD Interaction Graphics Display 交互式图形显示IGL Interactive Graphics Library 交互式图形库IGMP Internet Group Management Protocol 因特网组管理协议IGP Interior Gateway Protocol 内部网关协议IGRP Interior Gateway Routing Protocol 内部网关路由协议IGS Information Group Separator 信息组分隔符IHDL Input Hardware Des criptive Language 输入硬件描述语言IHL Internet Header Length 因特网报头长度IHV Independent Hardware Vendor 独立硬件商II Image Information 图像信息IIA Interactive Instructional Authoring 交互式教学写作IIA Internet Image Appliance 网络影像家电IIAS Interactive Instructional Authoring System 交互式教学写作系统IIC Incoming International Center 入局国际中心IID Image Intensifier Device 图像增强设备IIIN Intelligent Integrated Information Network 智能综合信息网络IIP Interface Information Processor 接口信息处理器IIS Internet Information Server 因特网信息服务器IIS Internet Information Service 因特网信息服务IISP Interim Inter-switch Signaling Protocol 临时的交换机间的信令协议IITA Information Infrastructure Technology and Application 信息基础设施技术及应用IITF Information Infrastructure Task Force 信息基础设施任务组IKBS Intelligent Knowledge Based System 基于知识的智能系统IKE Internet Key Exchange 因特网密钥交换IL Insertion Loss 插入损耗ILC Intelligent Line Card 智能线路卡ILD Insertion Loss Deviation 插入损耗偏差ILEC Incumbent Local Exchange Carrier 在业的本地交换运营公司ILI Idle Line Indicating 空闲线路指示ILMI Integrated Local Management Interface 综合本地管理接口ILMI Interim Local Management Interface 临时本地管理接口ILSLA Injection Locked Semiconductor Laser Amplifier 注入锁定半导体激光放大器IM Image Mixing 图像混合IM Instant Messaging 即时传信IM Integrated Modem 集成式调制解调器IM Interface Module 接口模块IM Inverse Multiplexing 反向复用IMA Interactive Multimedia Association 交互式多媒体协会IMAP Interactive Mail Access Protocol 交互邮件访问协议IMAP Internet Messaging Access Protocol 因特网消息存取协议IMAP4 Internet Message Access Protocol 4 因特网信息存取协议第4版IMC Inter-Module Connector 模块间连接器IMC International Maintenance Center 国际维护中心IMCC Inter-Module Communication Controller 模块间通信控制器IMEI International Mobile Equipment Identity 国际移动设备标识IMF InterMediate Fiber 中间光纤IMIS Integrated Management Information System 综合管理信息系统IMNI Internal Multimedia Network Infrastructure 多媒体网络内部基础设施IMP Interface Message Processor 接口报文处理器IMP Interface Module Processor 接口模块处理器IMS Information Management System 信息管理系统IMS Interactive Multimedia Service 交互式多媒体服务IMSI International Mobile Subscriber Identifier 国际移动用户标识符IMT Intelligent Multimode Terminal 智能多模式终端IMTC Internatinal Multimedia Television Committee 国际多媒体电视委员会IMTS Improved Mobile Telephone Service 改进的移动电话业务IMTV Interactive Multimedia TeleVision 交互式多媒体电视IMUX Input MUltipleX 输入复用IN Integrated Netowrk 综合网络IN Intelligent Network 智能网IN Interconnected Network 互连网络IN Internal Node 内节点IN-SL IN Service Logic 智能网业务逻辑IN-SM Intelligent Network Switching Manager 智能网交换管理器IN-SSM Intelligent Network Switching State Manager 智能网交换状态管理器IN-SSM Intelligent Network Switching Status Model 智能网交换状态模型INA Information Network Architecture 信息网体系结构INA Integral Network Arrangement 整体网络布局INA Integrated Network Architecture 综合网络体系结构INAP Intelligent Network Application Part 智能网应用部分INAP Intelligent Network Application Protocol 智能网应用协议INC Integrated Network Connection 综合网络连接INCC International Network Controlling Center 国际网络控制中心INCM Intelligent Network Conceptual Model 智能网概念模型INCS-1 Intelligent Network Capability Set-1 智能网能力组1INDB Intelligent Network DataBase 智能网数据库INDBMS Intelligent Network DataBase Management System 智能网数据库管理系统INE Intelligent Network Element 智能网元素INFM Intelligent Nework Functional Model 智能网功能模型INFO Integrated Network using Fiber Optics 采用光纤的综合网INI Intelligence Network Interface 智能网络接口INIC ISDN Network Identification Code ISDN网标识码INM Integrated Network Management 综合网络管理INM Integrated Network Monitoring 综合网监视INMARSAT INternational MARritime SAT ellite organization 国际海事卫星组织INMC International Network Management Center 国际网络管理中心INMOS IN service Management and Operation System 智能网业务管理及运行系统INMS Integrated Network Management System 综合网络管理系统INMS Intelligent Network Management System 智能网络管理系统INN Intermediate Network Node 中间网络节点INNO IN Network Operator 智能网运营商INP Intelligent Network Processor 智能网络处理器INS Information Network System 信息网络系统INS Intelligent Network Service 智能网络服务INSAT INternational SATellite 国际卫星INSES IN Services Emulation System 智能网业务仿真系统INSOS IN Service Operation System 智能网业务*作系统INSP Intelligent Network Service Provider 智能网服务供应商INSS Intelligent Network Service Subscriber 智能网业务用户INSTS IN Surveillance and Testing System 智能网监视和测试系统INT Interactive News Television 交互式电视新闻INTB IN TestBed 智能网试验台INTCO INT ernational COde of signal 国际信号码INTELSAT INternational TELecommunication SATellite 国际通信卫星(组织) INTIP INT egrated Information Processing 综合信息处理INTS Integrated National Telecommunication System 国家综合电信系统INTS INTernational Switch 国际交换INTS Inter-Network Time Slot 网络内部时隙INTSE INTelligent System Environment 智能系统环境IO Integrated Optics 集成光学IOAS Intelligence Office Automatic System 智能办公室自动化系统IOBB Input Output BroadBand 宽带输入输出IOC Input.Output Channel 输入/输出信道IOC Input / Output Controller 输入/输出控制器IOC Integrated Optical Circuit 集成光路IOC INTELSAT Operations Center 国际卫星组织*作中心IOC InterOffice Channel 局间信道IOCA Image Object Content Architecture 图像对象内容体系结构IOD Information On Demand 信息点播IODC International Operator Direct Calling 国际运营商直接呼叫IOLA Input / Output Link Adapter 输入/输出链路适配器IOLC Input / Output Link Control 输入/输出链路控制IOM Image-Oriented Memory 面向图像的存储器IOM Input / Output Multiplexer 输入/输出多路转换器IOM Integrated-Optic Modulator 集成光学调制器ION Integrated On-demand Network 综合按需服务网络IONI ISDN Optical Network Interface ISDN光网络接口IOP Input / Output Processor 输入输出处理器IOPDS Integrated-Optic Position / Displacement Sensor 集成光学位置/位移传感器IOS Integrated Office System 集成办公室系统IOS Intelligent Office System 智能办公室系统IOS Interactive Operating System 交互式*作系统IOS Internet Operating System 因特网*作系统IOS Internetwork Operating System 网间*作系统IOSB Input / Output Status Block 输入/输出状态块IOSC Input / Output Switching Channel 输入/输出交换通道IOSN Intelligent Optical Shuttle Node 智能光信息往返节点IOT Intra Office Trunk 局内中继IOTB Input / Output Transfer Block 输入/输出传送块IP Image Processing 图像处理IP Information Processing 信息处理IP Intelligent Peripheral 智能外设IP Internet Protocol 因特网协议IP Internetwork Protocol 网际协议IP Internetworking Protocol 组网协议IP Interworking Protocol 互通协议IPA Image Processing Algorithm 图像处理算法IPA Inerworking by Port Access 端口接入的互通IPBX International PBX 国际PBXIPC Integrated Peripheral Channel 集成外围通道IPC Intelligent Peripheral Controller 智能外设控制器IPC Inter-Personal Communications 人际通信IPC Inter-Process Communication 进程间通信IPC Inter-Processor Communication 处理器间通信IPCDN IP over Cable Data Network 电缆数据网传送IPIPCE International Path Core Element 国际通路核心单元IPCP IP Control Protocol IP控制协议IPCSM Input Port Controller SubModule 输入端口控制器的子模块IPDC IP Device Control IP设备控制IPE In-band Parameter Exchange 带内参数交换IPEI International Portable Equipment Identity 国际便携式设备标识IPF Image Processing Facility 图像处理设备IPG Interactive Program Guide 交互式节目指南IPG Inter-Packet Gap 分组信息间隙IPI Initial Protocol Identifier 初始协议标识符IPI Intelligent Peripheral Interface 智能外围接口IPL Initial Program Load 初始程序装入IPLB IP Load Balancing IP负载平衡IPLC International Public Leased Circuit 国际公用出租线路IPLI Internet Private Line Interface 因特网专用线接口IPLTC International Private Leased Telecommunication Circuit 国际专用租线通信电路IPM Inter-Personal Messeging 人际传信IPM-EOS Inter-Personal Message Element Of Service 人际报文业务单元IPM-UA Inter-Personal Messeging User Agent 人际传信用户代理IPME Inter-Personal Messaging Environment 人际传信环境IPMS Inter-Personal Messaging Service 人际传信业务IPMS Inter-Personal Messaging System 人际传信系统IPMS-MS Inter-Personal Messaging System Message Store 人际传信系统信息存储IPMS-UA Inter-Personal Messaging System User Agent 人际传信系统用户代理IPN Instant Private Network 瞬时专用网络IPN Inter-Personal Notification 人际通知IPng Internet Protocol next generation 下一代因特网协议IPOA IP Over ATM ATM网络承载IPIPP Internet Payment Provider 因特网支付业务提供商IPP Internet Platform Provider 因特网平台供应商IPPR Image Processing and Pattern Recognition 图像处理和模式识别IPR Intellectual Property Rights 知识产权IPS Image Processing System 图像处理系统IPS Information Processing System 信息处理系统IPS Information Protection System 信息保护系统IPS Intelligent Protection Switching 智能保护交换IPsec IP security protocol IP安全协议IPSF IP Service Function IP业务功能IPSS International Packet Switched Service 国际分组交换业务IPT Information Processing Technique 信息处理技术IPT Information Providing Terminal 信息提供终端IPUI International Portable User Identity 国际便携式用户标识IPv6 IP version 6 第六版IPIPX Internet Packet eXchange 因特网分组交换IPX Internetwork Packet eXchange 网际包(分组)交换IPX Interprocess Packet eXchange 进程间分组交换IQ Information Query 信息查询IQL Interactive Query Language 交互式查询语言IR Incoming Route 入路由IR Information Retrieval 信息检索IR InfraRed 红外IR Intelligent Robot 智能机器人IR Internal Router 内部路由器IRC Internet Relay Chat 因特网中继交谈IrDA Infra-red Data Association 红外数据协会IRFU Integrated Radio Frequency Unit 综合无线电频率单位IRI InfraRed Image 红外图像IRIM InfreRed Interface Module 远端接口模块IRIS Integrated platform for Regional Information System 地区信息系统用综合平台IRL Inter-Repeater Link 中继器间链路IRLAP InfraRed Link Access Protocol 红外链接存取协议IrLAP IrDA Link Access Protocol IrDA链路接入协议IRM Integrated Reference Model 综合参考模型IrMC Infrared Mobile Communication 红外移动通信IRN Information Resource Network 信息资源网络IRN Intermediate Routing Node 中间路由选择节点IRP Internal Reference Point 内部参考点IRP International Routing Plan 国际路由规划IRQ Information Repeat reQuest 信息重传请求IRSG Internet Research Steering Group 因特网研究指导组IRSU ISDN Remote Subscriber Unit ISDN远端用户单元IRTF Internet Research T ask Force 因特网研究任务工作组IS Imaging System 成像系统IS Information Science 信息科学IS Information System 信息系统IS Integrated Service 综合业务IS Intelligence System 智能系统IS Interactive Service 交互式业务IS Interactive Signal 交互信号IS Interface Specification 接口规范IS Interim Standard 临时标准IS-IS Intermediate System-to-Intermediate System 中间系统到中间系统ISA Industry Standard Architecture 工业标准体系结构ISA Information System Architecture 信息系统结构ISA Interim Standard Architecture 临时标准体系ISAN Integrated Service Analog Network 综合业务模拟网ISAP Interactive Speech Application Platform 交互语言应用平台ISAPI Internet Server Application Programming Interface 因特网服务器应用编程接口ISB Intelligent Signaling Bus 智能信令总线ISB Interface Schduling Block 接口调度块ISC International Switching Center 国际交换中心ISC Internet Software Consortium 因特网软件联盟ISC InterStellar Communications 星际通信ISCC International Service Coordination Center 国际业务协调中心ISCCI International Standard Commerical Code for Indexing 国际标准商用索引代码ISCII International Standard Code for Information Interchange 国际标准信息交换代码ISCP ISDN Signaling Control Part ISDN信令控制部分ISDCN Integrated Service Digital Center Network 综合业务数字中心网ISDN Integrated Service Digital Network 综合业务数字网ISDN-BA ISDN Basic rate Access ISDN基本速率接入ISDN-BRI ISDN Basid Rate Interface ISDN基本速率接口ISDN-PRA ISDN Primary Rate Access ISDN一次群速率接入ISDN-PRM ISDN Protocol Reference Model ISDN协议参考模型ISDN-SN ISDN Subscriber NumberISDN-UP ISDN User Part ISDN用户部分ISDS Integrated Switched Data Service 综合交换数据业务ISDT Integrated Service Digital Terminal 综合业务数字终端ISDX Integrated Service Digital eXchange 综合业务数字交换ISE Integrated Service Exchange 综合业务交换局ISE Integrated Switch Element 综合交换单元ISE Intelligent Synthesis Environment 智能综合环境ISEC Internet Service and Electronic Commerce 因特网服务和电子商务ISH Information Super Highway 信息高速公路ISIDE Interactive Satellite Integrated Data Exchange 交互式卫星综合数据交换ISL Inter-Satellite Link 卫星之间的链路ISLAN Integrated Services Local Area Network 综合业务局域网ISM Intelligent Synchronous Multiplexer 智能同步复用器ISM Interactive Storage Media 交互式存储媒体ISM Interface Subscriber Module 用户接口模块ISM Internet Server Manager 因特网服务器管理器ISM Internet Service Manager 因特网服务器管理程序ISMA Idle Signal Multiple Access 空闲信号多址ISMAN Integrated Services Metropolitan Area Network 综合业务城域网ISMC International Switching Maintenance Center 国际交换维护中心ISMS Image Store Management System 图像存储管理系统ISN Information System Network 信息系统网络ISN Integrated Services Network 综合业务网ISN Integrated Synchronous Network 综合同步网ISN International Signaling Network 国际信令网ISN Internet Shopping Network 因特网购物网络ISN Internet Support Node 因特网支持节点ISO International Standardization Organization 国际标准化组织ISOC Internet SOCiety 因特网学会ISODE ISO Development Environment ISO开发环境ISP Interactive Session Protocol 交互式会晤协议ISP Intermediate Service Part 中间业务部分ISP International Signaling Point 国际信令点ISP International Standardized Profile 国际标准化规格ISP Internet Service Provider 因特网服务供应商ISP Interoperable Systems Project 可互*作系统计划ISPBX Integrated Services PBX 综合业务PBXISPC International Signaling Point Code 国际信令点码ISR Initial Submission Rate 初始提供速率ISR International Simple Resell 国际简单转卖ISR Interrupt Service Routine 中断服务程序ISSLL Integrated Services over Specific Link Layer 专用链路层上的综合业务ISSS Interactive Subscriber Service Subsystem 交互式服务子系统ISTC International Satellite Transmission Center 国际卫星传输中心ISTC International Switching and Testing Center 国际交换和测试中心ISTV Integrated Service T eleVision 综合业务广播电视ISU Idle Signal Unit 空闲信号单元ISU Isochronous Slot Utilization 等时隙利用ISUP ISDN User Part ISDN用户部分ISV Independent Software Vendor 独立软件销售商ISVR Inter Smart Video Recorder 灵巧型视频录像机IT Information Technology 信息技术IT Information Theory 信息论IT International Transit 国际转接ITA International Telegraph Alphabet 国际电报字母表ITA2 International Telegraph Alphabet No.2 国际电报字母表第二版ITC Information Transfer Channel 信息传递信道ITC Intelligent Terminal Controller 智能终端控制器ITC International Telecommunication Center 国际电信中心ITC International Telephone Center 国际电话中心ITC International Television Center 国际电视中心ITC International Transit Center 国际转接中心ITC International Transmission Center 国际传输中心ITC InterToll Communication 长途局间通信ITCC International Telecommunication Control Center 国际电信控制中心ITD Interaural Time Difference 声源到达听者两耳的时间差ITDM Intelligent Time-Division Multiplexer 智能时分多路复用ITE Information Technology Equipment 信息技术设备ITE International Telephone Exchange 国际电话交换台ITF Information Transport Function 信息传送功能ITM ISDN Trunk Module 综合业务数字网中继模块ITMC International Transmission Maintenance Center 国际传输维护中心ITN Integrated Teleprocessing Network 综合远程处理网络ITN Intelligent Telecommunication Node 智能电信节点ITPC International Television Program Center 国际电视节目中心ITR Instantaneous Transmission Rate 瞬时传输速率ITR Internet Talk Radio 因特网无线对话ITS Independent Television Service 独立电视服务ITS Information Transfer System 信息转换系统ITS Information Transmission System 信息传输系统ITS Insertion Test Signal 插入测试信号ITS Intelligent Transport System 智能交通系统ITS International Telecommunication Service 国际电信业务ITSC International Telephone Service Center 国际电话业务中心ITSO International Telecommunications Satellite Organization 国际电信卫星组织ITSP Internet Telephony Service Provider IP电话业务提供商ITT InterToll Trunk 长途电话中继线ITTP Intelligent Terminal Transfer Protocol 智能终端转换协议ITTS Intelligent Target Tracking System 智能目标跟踪系统ITU International Telecommunication Union 国际电信联盟ITU-R ITU-Radio communications sector 国际电信联盟无线电通信组ITU-T ITU-Telecommunication standardization sector 国际电信联盟电信标准化组ITV Interactive TeleVision 交互式电视IU Interface Unit 接口单元IUI Intelligent User Interface 智能用户接口IUI Inter-User Interference 用户间干扰IUO Intelligent Underlay Overlay 智能双层网IUR Internet Usage Record 因特网使用记录IV Interactive Video 交互式视频IV Interface Vector 接口向量IVA Iitial Video Address 初始视频地址IVANS Insurance value Added Network Services 保险业增值网络服务IVAP Internal Videotex Application Provider 内部可视图文应用供应商IVBC International Videoconference Booking Center 国际电视会议登记中心IVC Independent Virtual Channel 独立虚拟信道IVC International Videoconference Center 国际会议电视中心IVD Interactive Video Disk 交互式视频盘IVD Interpolated Voice Data 内插语音数据IVDS Interactive Video Database Services 交互式视频数据库业务IVDT Integrated Voice Data Terminal 综合话音数据终端IVE International Videotex Equipment 国际可视图文设备IVG Interactive Video Game 交互式视频游戏IVHS Intelligent Vehicle and Highway System 智能车辆和公路系统IVIS Interactive Video Information System 交互视频信息系统IVMS Integrated Voice-Messaging System 综合语音信息系统IVN Interactive Video Network 交互式视频网络IVOD Interactive Video On Demand 交互式视频点播IVOT Inter-network teleVOTing 网间电子投票业务IVPN International Virtual Private Network 国际虚拟专用网IVR Integrated Voice Response 综合语音响应IVR Interactive Voice Response 交互式语音应答IVS Intelligent Video Smoother 智能视频平滑器IVS Interactive Video Service 交互式视频业务IVS Interactive Videodisc System 交互式视盘系统IW Information War 信息战IWAN Integrated services Wireless-Access Network 综合业务无线接入网络IWC Indoor Wireless Channel 室内无线信道IWC Interferometric all-optical Wavelength Converter 干涉全光波长变换器IWC Interferometric Wavelength Converter 干涉波长变换器IWCS Integrated Wideband Communication System 综合宽带通信系统IWF InterWorking Facility 互通设备IWF InterWorking Function 互通功能IWK Issuer Working Key 发行卡的工作密钥IWS Intelligent Work Station 智能工作站IWS Intelligent Workstation Support 智能工作站支持IWU InterWorking Unit 互通单元IXC Inter-eXchange Carrier 交互运营商IXP Internet eXchange Point 因特网交换点。

【英语教学法课件】Unit1Languageandlanguagelearning

3. Richards, J. & Rogers, T.S. Approaches and Methods in Language Teaching. (《语言教学(jiāo xué)路子与方法》1986)
4. Howatt, A.P.R. A History of English Language Teaching第十.五页(，共《78页。英语语言教学(jiāo
Audiolingualism
第二十五页，共78页。
听说法 (shuōf ǎ)
Audio-Lingual Method
❖ ‘Listen and repeat’ drilling activities are the most important classroom activities.
❖ Mistakes are immediately corrected and correct utterances are immediately praised.
❖ Functional view– communicative categories, communicative ability (to be able to communicate)
❖ Interactional view– to communicate appropriately (communicative strategies, cultural awareness, etc.)
❖ Language is a rule-based system and with a knowledge of the finite rules (language competence), infinite sentences can be produced

认知语言学流——入门必读起源,发展史,学术观点,研究方法

认知语言学认知语言学作为一种新的语言研究范式，产生于20世纪70年代，成熟于80年代，其标志是1989年在德国第一届国际认知语言学大会的召开和《认知语言学》的创刊。

其奠基性人物有G. Lakoff, R. Langacker, C. Fillmore, L. Talmy, M. Johnson, M. Turner, W.Chafe。

后来J.R. Taylor, D. Geeraerts, G. Fauconnier, E.Sweetser, A.Goldberg等人对认知语言学的发展做出了开创性贡献。

认知语言学是现代语言学中一个相当广泛的理论运动（movement）的总称。

它包含许多不同的途径、方法、研究重点。

这些不同的途径、方法和重点由共同的理论假设统一起来。

其中最重要的假设是语言是人类认知不可分离的一部分，任何对语言现象的真知灼见的分析都必须包含在人类认知能力之中。

认知语言学不能取代其他语言学理论或流派，相反，它与其他流派是互补的。

第一节认知语言学的产生、研究目标与语言观1.1 认知语言学产生的理论动因1.1.1 认知语言学产生的学术背景认知语言学属于认知科学的重要组成部分，产生于第二代认知科学（其主要理论观点将在1.3中讨论）。

第一代认知科学在认识论与方法论上具有以下特征。

第一，符号计算。

它认为理智(reason)与体验分离(disembodied)，是直义或客观的(literal), 就像形式逻辑一样是符号系统的运算。

因此，心智就是一个抽象的计算程序，心智的硬件（大脑和身体）对心智没有影响。

第二，意义就是心理表征。

可以作两种理解。

首先，意义根据符号之间的内在关系定义，意义就是符号计算的结果；表征即概念。

其次符号是外在现实的内在表征，即意义对应于客观现实；那么，表征就是形式系统之外某物的符号表征。

概而言之，思想可以用形式符号系统表征，而符号本身是没有意义的，思想是这些符号根据规则计算的结果。

ai advantage英语作文

ai advantage英语作文AI AdvantageIn recent years, the development of artificial intelligence (AI) has revolutionized the way we live and work. AI has rapidly integrated into various industries, providing numerous benefits and advantages. In this essay, we will explore the advantages of AI and how it is shaping our future.One of the main advantages of AI is its ability to improve efficiency and productivity. AI-powered systems can automate repetitive tasks, such as data entry, analysis, and customer service, allowing employees to focus on more strategic and creative work. This not only saves time but also increases accuracy and reduces human error. For example, AI algorithms can process large amounts of data in seconds, enabling companies to make faster and more informed decisions.Another advantage of AI is its ability to personalize experiences for users. AI-powered algorithms can analyze user data and behavior to provide personalized recommendations, content, and services. This customization enhances the user experience, increases engagement, and fosters customer loyalty. For example, AI-driven recommendation engines one-commerce websites can suggest products based on a user's browsing history and preferences, leading to higher conversion rates.AI also has the advantage of enhancing safety and security. AI-powered systems can detect patterns and anomalies in data to identify potential risks and threats. For example, AI algorithms can analyze video footage to recognize suspicious behavior and alert security personnel. Moreover, AI can be used to improve cybersecurity by detecting and mitigating cyber attacks inreal-time. By leveraging AI technology, organizations can better protect their assets and data from external threats.Furthermore, AI can improve healthcare outcomes and patient care. AI-powered systems can analyze medical data, such as images, genetic information, and patient records, to diagnose diseases, predict outcomes, and recommend treatment plans. This enables healthcare providers to deliver more accurate and personalized care to patients. For example, AI algorithms can interpret medical images, such as X-rays and MRIs, with higher accuracy than human radiologists, leading to faster and more reliable diagnoses.In addition, AI can drive innovation and create new opportunities for businesses. AI-powered technologies, such asmachine learning, natural language processing, and computer vision, can enable companies to develop new products, services, and business models. For example, AI chatbots can provide 24/7 customer support, AI-powered predictive analytics can optimize supply chain management, and AI-driven recommendation engines can personalize marketing campaigns. By embracing AI, organizations can gain a competitive edge and stay ahead of the curve.In conclusion, AI offers numerous advantages across various industries, including improved efficiency, personalized experiences, enhanced safety and security, better healthcare outcomes, and increased innovation. As AI continues to advance and evolve, its potential to transform our society and economy is limitless. By harnessing the power of AI, we can unlock new possibilities, drive growth, and create a better future for all.。

英语作文-人工智能助力电商平台推荐系统

英语作文-人工智能助力电商平台推荐系统Artificial Intelligence (AI) has significantly transformed various industries, and one of the areas where its impact is profound is in enhancing e-commerce platforms through recommendation systems. These systems play a crucial role in helping online platforms recommend products and services tailored to individual preferences, thereby improving user experience and boosting sales.E-commerce platforms rely heavily on recommendation systems to personalize the shopping experience for each user. Traditionally, these systems used basic algorithms like collaborative filtering or content-based filtering to suggest products based on purchase history or product attributes. While effective to some extent, these approaches have limitations in accurately predicting user preferences, especially in the context of rapidly changing consumer behavior and preferences.Enter artificial intelligence and machine learning. AI has revolutionized recommendation systems by enabling platforms to analyze vast amounts of data beyond just purchase history and product attributes. Machine learning algorithms can now process data on user behavior, including browsing patterns, search history, time spent on each product page, mouse movements, and even demographic information. This holistic approach provides a more comprehensive understanding of user preferences and intentions.One of the key advancements AI brings to e-commerce recommendation systems is its ability to perform real-time analysis. Unlike static rule-based systems, AI can continuously learn and adapt to new data, ensuring that recommendations remain relevant and up-to-date. This dynamic adaptation is crucial in fast-paced industries where trends can change rapidly, such as fashion and electronics.Moreover, AI-powered recommendation systems can identify subtle patterns in user behavior that human analysts may overlook. For example, AI algorithms can detect correlations between seemingly unrelated products that co-occur in user purchases or browsing sessions. This capability allows platforms to offer cross-selling opportunities,where complementary products are recommended together, thereby increasing the average order value.Another significant advantage of AI is its ability to personalize recommendations at scale. Whether a platform has thousands or millions of users, AI can create unique profiles for each individual based on their interactions with the platform. This personalization goes beyond just suggesting products; it extends to the presentation and timing of recommendations. For instance, AI can determine the optimal placement of recommendations on a webpage or tailor recommendations based on the user's current browsing session.Furthermore, AI enhances the accuracy of recommendations by integrating multiple sources of data. Beyond user interactions, AI can incorporate external factors such as seasonality, trends in social media, and even weather patterns (for certain types of products like clothing or outdoor gear). By considering these contextual factors, recommendation systems can offer more relevant suggestions that align with the user's current needs and preferences.Ethical considerations also play a crucial role in the development and deployment of AI in recommendation systems. As AI algorithms influence consumer choices, there is a responsibility to ensure transparency and fairness. Platforms must be transparent about how recommendations are generated and give users control over their preferences and data privacy. Moreover, AI systems should be designed to avoid reinforcing biases or promoting harmful content, ensuring that recommendations are beneficial and respectful to all users.Looking ahead, the future of AI in e-commerce recommendation systems holds promise for further advancements. As AI technology continues to evolve, we can expect more sophisticated algorithms that enhance personalization, improve prediction accuracy, and adapt to emerging consumer behaviors in real-time. Additionally, advancements in natural language processing and computer vision will likely expand the scope of AI applications in understanding and recommending products based on textual descriptions or image recognition.In conclusion, artificial intelligence has revolutionized e-commerce recommendation systems by enabling platforms to offer personalized and relevant suggestions to users at scale. Through advanced machine learning techniques, AI enhances accuracy, adapts in real-time, and considers a wide range of data sources to optimize user experience and drive business growth. As AI technology advances, the future of e-commerce recommendation systems looks promising, with continued improvements in personalization and user engagement.。

计算机视觉技术在自然语言处理中的使用方法

计算机视觉技术在自然语言处理中的使用方法自然语言处理（Natural Language Processing, NLP）是人工智能领域中重要的研究方向之一，旨在让计算机能够理解、分析和生成自然语言。

而计算机视觉技术（Computer Vision）则专注于使计算机能够从图像和视频中理解和获取信息。

这两个领域的结合为解决自然语言处理问题提供了更全面的方法。

计算机视觉技术在自然语言处理中的应用可以大致分为以下几个方面：1. 图像标注与描述生成图像标注是指给定一张图像，生成相应的文字描述。

通过结合计算机视觉技术的图像理解能力和自然语言处理的语义分析能力，可以让计算机生成更加准确和详细的图像描述。

这对于图像搜索、图像检索和辅助视觉障碍人士等应用有重要意义。

2. 视觉问答系统视觉问答系统是指根据给定的图像和提出的问题，通过理解图像内容并生成合适的自然语言回答来解答用户的问题。

计算机通过将图像转化为特征向量，然后与问题进行匹配，得到问题的答案。

这种系统结合了计算机视觉技术和自然语言处理技术，为用户提供了一种直观、便捷的交互方式。

3. 文字图像转换文字图像转换是指将具有文字内容的图像转换成为可供计算机进行自然语言处理的文本数据。

通过使用计算机视觉技术对图像中的文字进行识别和提取，可以使得文本数据能够进行更深层次的自然语言处理，如文本分类、情感分析等。

4. 视觉场景理解视觉场景理解是指使计算机能够从图像中识别和理解不同的视觉场景，如目标检测、物体识别和图像分割等任务。

这些任务的结果可以进一步用于自然语言处理任务，如自动图像字幕生成和智能图像搜索等。

5. 文本和图像的关联分析在某些应用中，文本和图像之间存在着紧密的关联。

例如，商品评论通常伴随着商品图片，新闻文章可能包含相关的图像等。

计算机视觉技术可以帮助自然语言处理系统更好地理解文本和图片之间的关系，提高文本理解的准确性。

综上所述，计算机视觉技术在自然语言处理中发挥着重要的作用。

随着科技的发展人工智能英语作文

随着科技的发展人工智能英语作文全文共3篇示例，供读者参考篇1The Development of Artificial Intelligence with Technological AdvancementsAs a student in the 21st century, it's impossible to ignore the rapid pace of technological advancements, particularly in the field of artificial intelligence (AI). The concept of AI, once confined to the realms of science fiction, has now become an integral part of our daily lives. From virtual assistants like Siri and Alexa to self-driving cars and intelligent robotics, AI has infiltrated nearly every aspect of our existence. As we stand on the precipice of a technological revolution, it's crucial to understand the implications and potential of this transformative technology.At its core, AI is the simulation of human intelligence processes by machines, particularly computer systems. These systems are designed to mimic human cognitive functions, such as learning, problem-solving, and decision-making. The field of AI encompasses a broad range of technologies, includingmachine learning, natural language processing, computer vision, and robotics, among others.One of the most significant drivers of AI development has been the exponential growth of computational power and the availability of vast amounts of data. Modern computers possess the ability to process and analyze enormous datasets, enabling them to identify patterns and make predictions with unprecedented accuracy. This has led to breakthroughs in areas such as image and speech recognition, language translation, and predictive analytics.Machine learning, a subset of AI, has been particularly instrumental in this progress. It involves training algorithms on vast amounts of data, allowing them to learn and improve over time without being explicitly programmed. This approach has revolutionized fields like recommendation systems, fraud detection, and predictive maintenance, among others.The impact of AI on our lives is already evident in countless ways. Virtual assistants like Siri and Alexa have become household names, helping us with tasks ranging from setting reminders to controlling smart home devices. Online shopping platforms leverage AI algorithms to provide personalized recommendations based on our browsing and purchase history.Social media networks employ AI to detect and remove harmful content, while email services use it to filter out spam and phishing attempts.Moreover, AI has made significant strides in the field of healthcare, aiding in disease diagnosis, drug discovery, and personalized treatment plans. In the realm of transportation, self-driving cars and autonomous vehicles are poised to revolutionize the way we commute, promising increased safety and efficiency on our roads.However, as with any transformative technology, the development of AI is not without its challenges and ethical considerations. One of the primary concerns is the potential for job displacement, as AI systems become increasingly capable of performing tasks traditionally carried out by humans. This has sparked debates around the need for reskilling and workforce adaptation to ensure a smooth transition into an AI-driven economy.Privacy and data security are also critical issues, as AI systems rely heavily on vast amounts of personal data for training and decision-making. There is a delicate balance to be struck between harnessing the power of data and protecting individual privacy rights.Furthermore, the development of AI raises questions about bias and fairness. AI algorithms can inadvertently perpetuate and amplify societal biases present in the data they are trained on, leading to discriminatory outcomes. Ensuring transparency, accountability, and ethical governance in the development and deployment of AI systems is paramount.Despite these challenges, the potential benefits of AI are too significant to ignore. As students, we stand at the forefront of this technological revolution, poised to shape and be shaped by the advancements in AI. Embracing this technology and developing the necessary skills to navigate this rapidly evolving landscape is crucial for our future success.Educational institutions play a pivotal role in preparing the next generation for an AI-driven world. Curricula should emphasize not only technical skills in areas like data science, machine learning, and programming but also critical thinking, ethical reasoning, and interdisciplinary collaboration. By fostering a well-rounded understanding of AI and its implications, we can cultivate a workforce capable of harnessing the power of this technology responsibly and ethically.Moreover, as students, we have a unique opportunity to shape the trajectory of AI development. By actively engaging inresearch, innovation, and entrepreneurship, we can contribute to the creation of AI solutions that address pressing societal challenges, from climate change to healthcare and beyond.In conclusion, the development of artificial intelligence is an unstoppable force, propelled by rapid technological advancements. As students, we stand at the forefront of this revolution, poised to witness and contribute to the transformative impact of AI on our lives. While the challenges are significant, the potential benefits are too vast to ignore. By embracing this technology with a critical and ethical mindset, we can harness the power of AI to create a better, more prosperous, and more equitable future for all.篇2With the Development of Technology, Artificial IntelligenceAs a student living in the 21st century, I can't help but be in awe of the rapid advancements in technology, particularly in the field of artificial intelligence (AI). It's a topic that has captivated my imagination and sparked countless debates among my peers and teachers. The notion of machines possessing human-like intelligence, once confined to the realms of science fiction, is now a reality that is shaping the world around us.From personal assistants like Siri and Alexa to self-driving cars and intelligent robots, AI has already woven its way into our daily lives. Its impact is undeniable, and its potential is vast. As students, we are witnessing firsthand how AI is transforming the educational landscape, offering new tools and resources that were unimaginable just a few years ago.One of the most exciting applications of AI in education is personalized learning. With the ability to analyze vast amounts of data and adapt to individual learning styles, AI-powered systems can tailor educational content and teaching methods to each student's unique needs and strengths. This approach has the potential to revolutionize the traditional one-size-fits-all model of education, ensuring that no student is left behind or held back.AI-assisted tutoring and adaptive learning platforms are already becoming commonplace in many classrooms. These systems can provide real-time feedback, identify areas where students are struggling, and adjust the pace and content accordingly. This level of personalization was previously unattainable, and it promises to bridge the gap between students who learn at different rates or have diverse learning preferences.Moreover, AI is opening up new avenues for research and discovery. By analyzing vast amounts of data and identifying patterns that may be invisible to the human eye, AI algorithms are aiding scientists in fields as diverse as medicine, astronomy, and environmental studies. As students, we are excited by the prospect of contributing to these groundbreaking endeavors, leveraging the power of AI to unlock new knowledge and solve complex problems.However, as with any transformative technology, the rise of AI also raises important ethical and societal concerns. One of the most pressing issues is the potential impact on employment. As AI systems become more advanced and capable of performing tasks traditionally reserved for humans, there is a legitimate fear of job displacement across various industries. This threat has sparked debates about the need for robust workforce retraining programs and the exploration of new economic models that can accommodate the changing landscape of work.Another crucial concern is the potential for AI to perpetuate or amplify existing biases and inequalities. AI systems are trained on vast datasets, and if those datasets reflect societal biases or lack diversity, the resulting models may reinforce harmful stereotypes or discriminate against certain groups. As students,we must grapple with these ethical dilemmas and advocate for the responsible development and deployment of AI technologies that prioritize fairness, transparency, and accountability.Despite these challenges, I remain optimistic about the transformative potential of AI. As a student, I am excited to be part of a generation that will shape the future of this technology and harness its power to address some of the most pressing challenges facing humanity.One area where AI could make a profound impact is in addressing global issues such as climate change, food insecurity, and disease prevention. By analyzing vast amounts of data and identifying patterns and trends, AI systems can aid in developing more effective strategies for mitigating environmental degradation, optimizing agricultural practices, and detecting early warning signs of disease outbreaks.Furthermore, AI has the potential to revolutionize fields like education, healthcare, and scientific research. Imagine a world where personalized learning is the norm, where medical diagnoses are more accurate and tailored to individual patients, and where breakthroughs in areas like renewable energy and sustainable development are accelerated by the power ofAI-driven analysis and problem-solving.However, realizing this potential will require a concerted effort from researchers, policymakers, and society as a whole. As students, we must engage in thoughtful discussions about the ethical implications of AI and advocate for responsible development and deployment practices. We must also ensure that AI technologies are accessible and beneficial to all, regardless of socioeconomic status or background.One way to achieve this is by promoting education and training in AI-related fields, fostering a diverse and inclusive workforce that can shape the development of these technologies. Additionally, we must encourage interdisciplinary collaboration, bringing together experts from various fields to tackle complex challenges and ensure that AI systems are designed with a holistic understanding of their potential impacts.As I look to the future, I am filled with a sense of wonder and excitement about the possibilities that AI holds. However, I also recognize the immense responsibility that comes with shaping this transformative technology. As students, we have the unique opportunity to be at the forefront of this revolution, to learn and grow alongside these advancements, and to contribute our perspectives and ideas to the ethical and responsible development of AI.It is a future filled with both challenges and opportunities, but one thing is certain – the age of artificial intelligence is upon us, and it will fundamentally change the way we live, learn, and understand the world around us. As students, it is our duty to embrace this revolution, to ask difficult questions, and to ensure that AI becomes a force for good, empowering humanity to reach new heights of knowledge, innovation, and progress.篇3The Rise of AI: How Technology is Reshaping Our WorldAs a student living in the digital age, I can't help but be in awe of the rapid advancements in artificial intelligence (AI) technology. It's a topic that both excites and unsettles me, as I ponder the implications it will have on our future. On one hand, AI holds the promise of revolutionizing various industries, streamlining processes, and solving complex problems that have long eluded us. On the other hand, the prospect of intelligent machines surpassing human capabilities raises ethical concerns and existential questions about our role in an increasingly automated world.Let's start with the remarkable progress AI has made in recent years. From voice assistants like Siri and Alexa toself-driving cars and sophisticated language models, AI has become an integral part of our daily lives, often without us even realizing it. The ability of machines to process vast amounts of data, learn patterns, and make decisions at lightning speed is nothing short of astonishing. In fields like healthcare, AI is being used to analyze medical images, identify potential diseases, and even assist in drug development, potentially saving countless lives.Moreover, AI has revolutionized the way we interact with technology. Natural language processing (NLP) has made it possible for us to communicate with machines using our native tongues, breaking down language barriers and making technology more accessible to people from diverse backgrounds. AI-powered translation tools have also facilitated cross-cultural communication and understanding, bringing the world closer together.However, as awe-inspiring as these advancements are, they also raise important questions about the ethics and societal impact of AI. One of the most significant concerns is the potential for job displacement as machines become capable of performing tasks traditionally done by humans. While some argue that AI will create new job opportunities in fields such asprogramming and data analysis, there is a legitimate fear that many professions, particularly those involving repetitive or routine tasks, could be automated, leading to widespread unemployment and economic disruption.Another pressing issue is the potential for AI to perpetuate or even amplify existing biases and discrimination. Since AI systems are trained on data created by humans, they can inherit and reinforce societal biases present in that data, leading to unfair and discriminatory outcomes. For instance, if anAI-powered hiring system is trained on data that reflects existing gender or racial biases in the workforce, it may continue to discriminate against certain groups, further exacerbating inequalities.Furthermore, the increasing reliance on AI raises concerns about privacy and data security. As AI systems become more sophisticated, they require access to vast amounts of personal data to function effectively. This data could potentially be misused, compromised, or even weaponized by malicious actors, posing a significant threat to individual privacy and security.Despite these concerns, I believe that the potential benefits of AI outweigh the risks, provided we approach its development with caution, ethical considerations, and robust regulatoryframeworks. One area where AI could have a profound impact is in addressing global challenges such as climate change, food insecurity, and disease outbreaks. By harnessing the power of machine learning and data analysis, AI could help us develop more sustainable and efficient solutions, predict and mitigate environmental risks, and accelerate scientific research and discovery.Moreover, AI has the potential to revolutionize education and make learning more personalized and accessible.AI-powered tutoring systems could adapt to individual learning styles and pace, providing tailored instruction and feedback to students. Virtual reality and augmented reality technologies, powered by AI, could create immersive and interactive learning experiences, making complex concepts more engaging and easier to understand.As a student, I am particularly excited about the potential of AI to enhance creativity and artistic expression. AI-powered tools could assist writers, musicians, and artists in generating ideas, exploring new styles, and pushing the boundaries of their creativity. While some may argue that AI could eventually replace human artists, I believe that true creativity and emotional expression will always remain a uniquely human endeavor, andAI will simply be a tool to augment and enhance our creative abilities.However, for AI to truly benefit humanity, we must address the ethical and societal concerns head-on. This requires a multifaceted approach involving policymakers, technologists, ethicists, and the general public. Robust regulations and guidelines must be put in place to ensure the responsible development and deployment of AI systems, with a focus on transparency, accountability, and the protection of individual rights and privacy.Additionally, we must prioritize education and awareness about AI, both in formal educational settings and in broader public discourse. It is crucial that people understand the strengths and limitations of AI, as well as its potential impacts on society. Only through informed and open discussions can we navigate the complexities of this technology and shape its development in a way that benefits all of humanity.Ultimately, the rise of AI is both exhilarating and daunting, but it is a reality we cannot ignore. As a student, I feel a sense of responsibility to engage with this technology, to understand its implications, and to be part of the conversations that will shape its future. While the path ahead is uncertain, I am confident thatwith the right approach, AI can be a powerful tool for solving some of the world's most pressing challenges and improving the human condition. It is up to us, the next generation, to harness the potential of AI while mitigating its risks, ensuring that it serves the greater good of society.。

给客户推销相机的英语作文

As a high school student with a keen interest in photography, Ive always been passionate about capturing the world around me. This passion led me to learn about different types of cameras and their capabilities. Recently, I had the opportunity to introduce a new camera model to a potential customer, and Id like to share my experience.It was a sunny Saturday afternoon when I walked into the local electronics store for a parttime job. The store was bustling with people looking for the latest gadgets. I spotted a middleaged man browsing through the camera section, seemingly overwhelmed by the variety of options. Sensing his hesitation, I approached him with a friendly smile.Excuse me, sir, I began, I noticed youre looking at cameras. Id be happy to help you find the perfect one for your needs.He looked relieved and nodded. Im trying to find a camera for my upcoming trip to the mountains. I want something that can capture the beauty of the scenery, but Im not sure which one to choose.I listened carefully to his requirements and decided to introduce him to our latest model, the Eagle Eye 360. This camera is designed for adventurers and nature lovers like him, offering highquality images and a wide range of features.The Eagle Eye 360 is equipped with a 24megapixel sensor, ensuring that every detail of the mountain landscape is captured with clarity, I explained. It also has a 3inch LCD screen, allowing you to easily frame your shots andreview your photos on the spot.The mans eyes lit up with interest, so I continued, One of the standout features of this camera is its 360degree panoramic mode. With just a simple sweep of the lens, you can capture the entire mountain vista in a single, stunning image. Imagine being able to share the breathtaking views with your friends and family as if they were standing right beside you.He seemed impressed but still had some concerns. That sounds great, but what about the battery life? Ill be out in the wilderness for days, and I dont want to worry about recharging.I smiled, confident in the cameras capabilities. The Eagle Eye 360 has an impressive battery life of up to 500 shots on a single charge. Plus, its compatible with standard AA batteries, so you can easily replace them if needed. You wont have to worry about running out of power during your trip.As we continued our conversation, I showed him the cameras other features, such as its builtin GPS for geotagging your photos, its weatherresistant body for protection against the elements, and its userfriendly interface for easy operation even for beginners.The man was clearly intrigued and asked to test the camera. I handed it to him, and he took a few test shots, marveling at the image quality and ease of use. After a few more questions and a comparison with other models, he finally decided to purchase the Eagle Eye 360.As he walked out of the store with his new camera, I felt a sense of accomplishment. Not only had I helped him find the perfect tool for his mountain adventure, but I had also gained valuable experience in sales and customer service.This experience taught me the importance of understanding the customers needs, showcasing the products features effectively, and addressing any concerns they might have. Its not just about selling a product its about creating a connection and ensuring that the customer is satisfied with their purchase.In conclusion, introducing a camera to a customer involves more than just listing its specifications. Its about understanding their needs, showcasing the cameras capabilities, and providing a solution that meets their requirements. As a high school student, this experience has not only enhanced my knowledge of cameras but also given me insights into the world of sales and customer relations.。

1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。

Image Browsing and Natural Language Paraphrases ofSemantic Web AnnotationsChristian Halaschek-Wiener, Jennifer Golbeck, Bijan Parsia,Vladimir Kolovski, and Jim HendlerMaryland Information and Network Dynamics Lab, 8400 Baltimore Ave.College Park, MD, 20740 USA{halasche,golbeck,kolovski,hendler}@bparsia@Abstract. Recently, there has been interest in marking up digital images withannotations describing the content of the images using Web based ontologiesencoded in the W3C’s Web Ontology Language, OWL. The annotations aresubsequently exploited to improve the user experience of large collections ofimages, whether by enhance search or by a structured browsing experience. Inthe latter case, the complexity and unfamiliarity of logic-based ontology lan-guages may do more to impede, than aid, the user. To alleviate this problem, wepropose using automatic generation of natural language (NL) paraphrases ofOWL statements to assist browsing image content. In this paper, we provide anoverview of our NL generation approach and an empirical evaluation of the useof our paraphrases for image browsing.1 IntroductionRecently there has been interest in using Semantic Web ontologies encoded in the Web Ontology Language [3], OWL, to formally represent the semantic content of digital images.[1]. Given semantically rich image metadata, collections can be more accurately searched and browsed, with new knowledge derived from existing annota-tions. Additionally, by exploiting standard Web mechanisms, one is able to link exist-ing image collections to arbitrary knowledge repositories and vice versa.The canonical OWL interchange syntax is based on the XML serialization of RDF (RDF/XML). Neither RDF nor XML were designed readability in mind (much less casual end user readability). There are several alternative surface syntaxes designed with more readability in mind [14, 15] and, due to the correspondence with first order logic, there are a number of traditional logic notations available. However, these all require a fairly deep understanding of OWL, logic, or both which, aside from being comparatively rare, seems unnecessary for the purpose of navigating through image collections. Recently, there has been work in providing natural language translations of OWL concept definitions, providing users with a format that is easier to both readand understand [2, 6, 8, 9]. In [2], we provided an algorithm for paraphrasing OWL concepts with a controlled natural language.In this paper, we propose the use of automatically generated NL paraphrase of OWL classes and individuals to aid in browsing image. We feel that the paraphrases will make allow users to more effectively use and enjoying using the semantic annota-tions when browsing image collections, and to better understand the content of the browsed images.Most semantic markup of images includes a large number of instance assertions, corresponding to features of concrete particular depicted in the images. Therefore, we extend the algorithm in [2] to provide natural language paraphrases of OWL asser-tions about individuals, as well as of class definitions. We have implemented this algorithm as a plug-in to an ontology-based, image annotation and browsing tool, PhotoStuff [1]. Lastly, we provide an empirical evaluation of the effectiveness of these NL paraphrases for the task of browsing and interacting with image collections.2 Background[2] presents an approach for automatically generating natural language paraphrases for OWL concept definitions that attempts to preserve the underlying meaning of the logic based definition. The approach is applicable to any ontology where the classes and properties are named using certain standard naming conventions. The algorithm is fairly straightforward considering the good quality results achieve. The most sophisti-cated NL processing tool the algorithm uses is a part-of-speech (POS) tagger, which is used to improve the fluency of the generated NL description.The first step in the approach is to generate a parse tree corresponding to the rela-tions between the OWL class and other entities. Creating the tree provides additional flexibility to alter it (post-process) in any way deemed necessary (a sample parse tree is provided in [2]).In [2], it has been noted that property names from several major ontologies reveal that, while properties could theoretically be named with arbitrary words, their names are generally parsed into one of a small number of simple phrase structures. These structures can algorithmically be restructured, by using a POS tagger, providing a (more) natural language style format. Below, Table 1 lists these phrase structure cate-gories, along with their reformatted natural language translations.Table 1. Common class and property phrase structure, along with natural language translationsAfter generating the parse tree, there are several steps in generating the NL output, including a pre-processing step where the tree is modified to eliminate nodes contain-ing owl:Thing, etc. Further details are available in [2]. In general, the approach gener-ates full English sentences whenever possible. However, we have found that render-ing complex concepts entirely in NL sometimes results in very lengthy, difficult to understand sentences. In some cases, using a bulleted, nested list format for such complex sets of conditions was much clearer. For example, part of the definition of Beaujolais from the wine ontology1 is given below in Table 2.Table 2. Bulletized natural language rendering of OWL class BeaujolaisA Beaujolais is a Wine that:- is made from at most 1 grape, which is Gamay Grape- has Delicate flavor- has Dry sugar- has Red color- has Light body1 Wine OWL Ontology: /2001/sw/WebOnt/guide-src/wine.rdf3 OWL Individual NL ParaphrasesWe can categorize images based on the types of things they depict. In the SemSpace2 portal, a nested list rendering of the class hierarchy is the initial interface to the col-lection. Thus, NL paraphrases of concepts might help users determine which catego-ries are likely to contain images of interests. However, images, especially photo-graphs, generally depict concrete objects that are naturally represented in OWL as individuals with various types and properties. Thus, if paraphrases are to help the user understand the contents of particular images, we must extend the algorithm in [2] to handle NL paraphrases of OWL individuals as well.3.1 Approach OverviewThe approach adopted here provides an NL rendering of OWL individuals based on the direct relations of that individual (e.g., type assertions, labels, defined relations, etc.). In order to generate an NL paraphrase for an OWL individual, first a NL parse tree is generated for that individual [2]. In contrast to the approach for OWL concepts [2], this tree will be at most one level deep, as only the direct relations are used. As the tree is created, only rdf:type assertions corresponding to the (possibly inferred) most specific classes of the individual are added as edges. (In certain ontologies with shallow, informative class graphs, it might be preferable to add all the types of the individuals, or to control the depth in a different way.) The relation edges are labeled by two distinct mechanisms: An rdf:type edge in the tree is given an “is a”label. For all additional relations, the POS tagger is run over the “local part” of the property’s URI (as in [2]) to provide more legible labels for the relations. Finally, the labels for the objects of the relations are added to the object nodes. Figure 1 depicts a subset of the NL parse tree and the original RDF/XML for a sample individual, Storey Mus-grave.Fig. 1. Subset of RDF/XML and of NL Parse Tree for OWL Individual Storey Musgrave2 /After this post-processing the NL rendering is generated. The current approach uses the following template for producing OWL individual NL renderings: First the label edges of the NL tree for the given individual are retrieved. This label is used to begin the NL sentence. Following this, a comma separated rendering of all rdf:type edges is generated. Then a bullet list of all the additional relations is produced. The template is shown below in Table 3.Table 3. RDF/XML serialization of individual Storey Musgrave vs. the automatically genera-ted natural language rendering. Italicized words are invariant of the individual being renderedIndividual is a rdf:type that- relation 1- relation 2- …- relation nFor example, the NL rendering for the OWL individual Storey Musgrave is pre-sented below in Table 4 (note that the original RDF/XML for Storey Musgrave is provided in Table 3).Table 4. Automatically generated natural language rendering of Storey Musgrave4 NL Captions for Browsing Image CollectionsIn this section we provide details of using automatically generated NL paraphrases for browsing image collections in our image annotation tool, PhotoStuff [1]. The main motivation for using NL paraphrases for browsing image collections and their annota-tions is that it will provide a more enjoyable user experience. Essentially, the user will be provided with the details about image contents in a human readable and under-standable format, which allows them to focus on the images and what they depict rather than the cryptic details of the representation language. One additional benefit of publishing the natural language rendering of the images is that it can allow exiting search engine technology to index such content.4.1Implementation DetailsPhotoStuff is a platform independent, open source, image annotation tool that allows users to annotate images and their sub-regions using concepts from any number of ontologies specified in OWL. PhotoStuff can also load pre-existing annotations, which can then be browsed or used in subsequent annotations.An experimental implementation of our NL generation algorithm has been pro-vided as a plug-in for PhotoStuff. Within PhotoStuff, the NL captions are used in two main ways. First, when an image is selected and loaded into the image canvas, NL renderings for the instances depicted within regions are provided as pop-ups when the regions are moused over (see Figure 2). This provides the user with a quick, human readable display of the individual depicted within that region.Fig. 2. Region NL Caption Pop-UpAdditionally, the NL for individuals depicted within regions can be viewed by right-clicking the regions, and selecting “View NL”. This puts the natural language paraphrases in a NL info pane, as shown below in Figure 3. The figure illustrates that in the NL info pane, the generated natural language includes hyperlinks for all addi-tional individuals and OWL class that occur in the rendering of the current individual or class. This allows the user to browse the image collection based on the existing annotations in a Web-like manner. Further, the hyperlinks can be right clicked, pro-viding the option to filter the thumbnail strip to only show images that depict that instance or instances of that class.Fig. 3. NL Info Pane for Browsing Image Annotations in NL Format5 EvaluationTo evaluate the benefits of the NL format, we conducted a pilot user study. Our hy-pothesis was that users would prefer the NL format for viewing data when the task was to gain an understanding of the meaning of a concept.Our pilot study included seven subjects. Subjects ranged in age from 20 to 37 and all were students working toward bachelors, masters, or Ph.D. degrees. We tested the subjects’ preferences for method of viewing classes separately from viewing in-stances. For both classes and instances, we choose three examples: one very simple, one of medium complexity, and one very complicated example.When viewing classes, we used the Wine ontology mentioned earlier. Our simple class is AlsatianWine, which only had one restriction. Anjou was the medium com-plexity class, with four simple restrictions and an intersection combined with a restric-tion. The most complex class was Beaujolais, with six simple restrictions, including cardinality, and an intersection combined with a restriction.Each class was presented to the subjects in SWOOP [13]. For the study, four addi-tional formats for viewing each class was provided in SWOOP (see Figure 4).Figure 4. Different view renderings used for evaluation of class descriptions. Clock-wise from upper left: Concise format, Turtle, Abstract Syntax, and RDF/XML.Users were asked to view the concept with the goal of understanding it’s meaning, rather than the modeling features. They were allowed to take as much time as they needed and, when finished, they were asked to rank the formats according to their preference, with 1 being best and 5 being worst. If the subjects felt two formats were equally good, they were allowed to give them the same ranking.For all three classes, the order of the rankings was the same. From best to worst, the formats were ranked, and shown below in Table 5.Table 5. Rankings from best to worst of view format of class descriptions1. Natural Language2. Concise Format3. Abstract Syntax4. Turtle5. RDF/XMLAdditionally, the average ranking and standard deviations for the various formats is presented below in Table 6.Table 6. Average rankings and standard deviations for the five formats used to dis-play class informationAverage Rank (Standard Deviation)Class ConciseFormat AbstractSyntaxNaturalLanguageRDF/XML TurtleAlsatianWine 2.00(1.41) 3.57(1.13) 1.57(0.79) 4.86(0.38) 3.86(0.69) Anjou 1.86(0.69) 2.71(0.49) 1.14(0.38) 5.00(0.00) 4.00(0.00) Beaujolias 2.14(0.69) 2.71(0.49) 1.00(0.00) 5.00(0.00) 3.86(0.38)Using a Wilcoxon matched-pairs signed-ranks test, we found that the NL format significantly outperformed the Concise Format (ranked second on average) for both Anjou and Beaujolais with p<0.05. There was not a significant benefit over the Con-cise format for the simplest class, AlsatianWine, but NL did significantly outperform the Abstract Syntax, which was the average third ranked format. This allows us to conclude that in the pilot study the NL format offers significant benefits to users when they are trying to understand the meaning of classes, particularly complex classes.To evaluate the NL format with instances, subjects were given three instances from the collection of astronaut data3. The first instance, Bryan O'Conner, had only one property: the depiction that tied him to an image. The second instance, John Young, had a depiction and one additional property. The last instance, Storey Musgrave, was tied to three regions and four properties, including both Datatype and Object proper-ties. Subjects viewed the data about each instance in a tabular format and in the NL format with the goal of understanding the information about each instance. Subjects took the time they needed without limits and then were asked to choose which format they preferred.For each of the three instances, the NL format was preferred 6:1 over the instance form. When subjects commented on their preference, it focused largely on the fact that the NL format was more concise, displaying only relevant information.We conclude this section by stating that the results of this pilot study show that us-ers feel that the NL format provides them with an advantage when trying to under-stand both classes and instances.6 Discussion and Future DirectionsWhile our initial evaluations support the usefulness of the approach presented here, we note there are a few limitations, which we leave as future work. First, if individu-als participate in a large number of relations, then the generated bullet list, or con-3 SemSpace portal instance data. Available at /rdf/dumpcatenated sentence, can be quite long and cumbersome to read through. One potential solution to ease this problem is to use a ranking metric when displaying the relations in the NL generated format. [4, 5] propose relationship-ranking metrics that could potentially be used to filter irrelevant relations, thus presenting the user with a smaller NL paraphrase.An additional solution to this problem we would like to investigate is the to utilize text summarization techniques for the output from the NL generation algorithm. [10, 11, 12] present a variety of techniques for summarizing such textual data. Solutions such as these could possibly be applied to the NL output, thus presenting the user with a succinct version of the output from the NL generation engine.In this work, we use a predefined template view of the NL rendering of instances. Under some circumstances, particular users may wish to define their own templates. Thus we would like to support using custom templates for particular ontologies or class descriptions.In our pilot study, we focused on user understanding of OWL data and found, gen-erally, that users preferred the NL paraphrases for that purpose. While we found this format to be preferred for understanding image annotations, the extent of which these paraphrases work to improve the overall image browsing experience remains to be seen. Therefore we plan to perform further evaluations of the approach.7 Related WorkIn this section we present related work in the area of automatically generating natural language translations of Semantic Web ontological concepts and instances. [6] pro-vides an instructional use of NL paraphrases for understanding OWL Concepts. The authors mention a plug-in to Protégé, the Class Description Display4plug-in that provides simpler NL descriptions that resemble OWL Abstract Syntax. This work has been refined into the so-called Manchester syntax.5 In [7], a subset of English is intro-duced called Attempto Controlled English (ACE). ACE is translated into first-order logic and thus can be used as a formal notation. Therefore, ACE is a formal language with the semantics of first order logic. In comparison, our approach converts OWL classes and individuals, which are based on a decidable subset of FOL called Descrip-tion Logics, into a NL description. [8] describes an XML-based NL generation for RDF and DAML+OIL. In this work, a pipeline of XSLT transformations implements the sequence of processing stages in the orthodox pipeline architecture for NL genera-tion. The generator uses predefined XSLT text plan templates for specific ontologies, following a domain-specific approach of shallow generation. However, it is unclear whether this approach works efficiently for more complex OWL ontologies. [9] pre-sents an approach for automatic generation of reports from OWL ontologies, using natural language generation tools. Our work here differs in that we do not rely on a lexicon for NL generation.4 Class Description Display plug-in. Available at /downloads/cdc/5 /resources/reference/manchester_syntax/7 ConclusionsIn this paper we have provided an extension of our previous work [2] to automatically generate natural language paraphrases for OWL individuals. Given this, we have proposed using this approach for browsing Semantic Web annotations of image col-lections. This provides the user with a more pleasant experience of the annotations.We have additionally presented an experimental implementation of the approach in an ontology based, image annotation tool, PhotoStuff. Finally, we have provided an empirical evaluation of the usage of the NL paraphrases for browsing image collec-tions, finding that it was in fact useful for the task.This work was supported in part by grants from Fujitsu, Lockheed Martin, NTT Corp., Kevric Corp., SAIC, the National Science Foundation, the National Geospa-tial-Intelligence Agency, DARPA, US Army Research Laboratory, and NIST. We would also like to thank Aditya Kalyanpur and Mike Grove for all of their contribu-tions to this work.References1. Halaschek-Wiener, C., Schain, A., Golbeck, J., Grove, M., Parsia, B., Hendler, J.: A FlexibleApproach for Managing Digital Images on the Semantic Web, 5th International Workshop on Knowledge Markup and Semantic Annotation, Galway, Ireland, November 7, 2005.2. Hewlett, D., Kalyanpur, A., Kolovski, V., Halaschek-Wiener, C.: Effective Natural Lan-guage Paraphrasing of Ontologies on the Semantic Web. End User Semantic Web Interac-tion Workshop, International Semantic Web Conference (ISWC), November 2005, Galway, Ireland.3. Dean, M., Schreiber, G.: OWL Web Ontology Language Reference. W3C Recommendation.February 2004. /TR/2004/REC-owl-semantics-20040210/4. Aleman-Meza, B., Halaschek-Wiener, ., Arpinar, B., Ramakrishnan, C., Sheth. A.: A Flexi-ble Approach for Analyzing and Ranking Complex Relationships on the Semantic Web, IEEE Internet Computing (Journal), Special Issue on Information Discovery: Needles and Haystacks, May-June 2005.5. Anyanwu, K., Maduko, A., Sheth, A.: SemRank: Ranking Complex Relationship SearchResults on the Semantic Web. The 14th International World Wide Web Conference, (WWW2005), Chiba, Japan, May 10-14, 20056. Rector, A., Drummond, N., Horridge, M., Rogers, J., Knublauch, H., Stevens, R., Wang, H.,Wroe, C.: OWL Pizzas: Common errors & common patterns from practical experience of teaching OWL, European Knowledge Acquisition Workshop (EKAW) 2004.7. Fuchs, N., Schwertel, U., Torge, S.: Controlled natural language can replace first-order logic.14th IEEE International Conference on Automated Software Engineering, Cocoa Beach, Florida.8. Wilcock, G.: Talking owls: Towards an ontology verbalizer. Proceedings of the 2nd Interna-tional Semantic Web Conference (ISWC), Oct. 2003.9. Bontcheva, K., Wilks, Y.: Automatic Report Generation from Ontologies: the MIAKT ap-proach, In Proceedings Nineth International Conference on Applications of Natural Lan-guage to Information Systems (NLDB'2004), Manchester, UK.10. Buyukkokten, O, Garcia-Molina, H., Paepcke, A.: "Seeing the Whole in Parts: Text Sum-marization for Web Browsing on Handheld Devices". The 10th International WWW Con-ference (WWW10). Hong Kong, China - May 1-5, 2001.11. Barzilay, R., Elhadad, M.: Using Lexical Chains for Text Summarization. In Proceedings ofthe ACL Workshop on Intelligent Scalable Text Summarization, pages 10–17, Madrid, Spain, August 1997.12. Hovy, E.H. Lin, C.: Automated Text Summarization in SUMMARIST. In M. Maybury andI. Mani (eds), Advances in Automatic Text Summarization. Cambridge: MIT Press. 1999.13. Aditya Kalyanpur, Bijan Parsia, Evren Sirin, Bernardo Cuenca-Grau, and James Hendler.Swoop - a web ontology editing browser. Journal of Web Semantics, 4(1), 2005.14. Beckett, D., Turtle — Terse RDF Triple Language,</2004/01/turtle/>, February, 2006.15. Patel-Schneider, P., Hayes, P., Horrocks, I., OWL Web Ontology Language Semantics and Abstract Syntax, </TR/owl-semantics/>, February, 2004.。