Haptic Interaction and Volume Modeling Tech-niques for Realistic Dental Simulation
Visual Comput(2006)22:90–98
DOI10.1007/s00371-006-0369-8O R I G I N A L A R T I C L E
Laehyun Kim Se Hyung Park Haptic interaction and volume modeling techniques for realistic dental simulation
Published online:31January2006
?Springer-Verlag2006
L.Kim(u)·S.H.Park
Korea Institute of Science and Technology, {laehyunk,sehyung}@kist.re.kr Abstract We present haptic simula-
tion and volume modeling techniques
for a virtual dental training system.
The system allows dental students to
learn dental procedures and master
their skills with realistic tactual
feelings.It supports various dental
procedures,such as dental probing,
to diagnose carious lesions,drilling
operation for cavity preparation,
and?lling the prepared cavities
with amalgam.The system requires
fast and stable haptic rendering and
volume modeling techniques work-
ing on the virtual tooth.Collision
detection and force computation are
implemented on an offset surface in
volumetric representation to simulate
reasonable physical interactions
between dental tools with a certain
volume and the teeth model.To avoid
discrete haptic feeling due to the
gap between the fast haptic process
(1KHz)and much slower visual
update frequency(30Hz)during
drilling and?lling the cavities,we
employed an intermediate implicit
surface to be animated between the
original and target surfaces.The
volumetric teeth model is converted
into a geometric model by an adaptive
polygonization method to maintain
sharp features in every visual frame.
V olumetric material properties are
represented by stiffness and color
values to simulate the resistance and
texture information depending on
anatomical tissues.Finally,we made
a dental workbench to register sen-
sory modalities like visual,auditory
and haptic sensation.
Keywords Dental simulation·
Haptic interaction·V olumetric
implicit surface·Dental workbench
1Introduction
Usually dental students use plastic teeth(called the ty-
podont model)working with real dental instruments to practice cavity preparation and other dental procedures.
However,plastic teeth lack the level of detail and mate-rial properties needed to accurately simulate real teeth and procedures.Recently,virtual reality(VR)dental training systems incorporating haptic interface have been intro-duced as alternatives to traditional training procedures.This allows students to perform virtual dental procedures by providing realistic tactile sensations in the same way that they do in the real world.
The dental training system using haptic interface re-quires two basic techniques:haptic rendering to generate force feedback and drilling simulation for the material cutting behavior from the virtual tooth model.Typically, the volume-based approach is used for drilling simulation, since it provides direct and intuitive modeling without topological constraints and volumetric material represen-tation,unlike the geometry-based one.
Haptic interaction and volume modeling techniques for realistic dental simulation 91
In recent years,many haptic modeling systems based on volumetric data have been proposed for virtual design,art,or surgical training.FreeForm TM [5]is the ?rst com-mercial haptic modeling system from SensAble.However,it is are not suitable for realistic dental simulation because the systems cannot simulate the various materials of the tooth,smooth force rendering,debris by the dental drill,or physical contact sound between the tools and the virtual teeth.
In this paper,we propose volume-based haptic dental simulation techniques to address limitations in previous systems as follows:
–Collision detection and force computation based on an offset surface for simulating realistic tool interaction with the virtual tooth.
–Smooth force rendering by directly simulating in-termediate implicit surfaces created by interpolating potential values between two consecutive physical models during drilling (carving operation)and ?lling the prepared cavity (adding operation).
–Haptic simulation of internal material properties,in-cluding stiffness and color information,which are saved into a volumetric representation.
–Multi-resolution mesh generation from the volume model being modi?ed using an adaptive polygoniza-tion method.The generated mesh is adapted to surface and color complexity,respectively,to maintain sharp features and clear color boundaries.
–Simulation of physical contact sound using the AR (auto regressive)model.
Our system consists of two basic parts:the visual and haptic processes run on a PC with dual Intel Xeon 3.0-GHz CPUs,2Gbyte RAM,and an NVidia FX3000video card.The visual rendering implementation (includ-ing simulation)is OpenGL-based and a 3-DOF
PHAN-
Fig.1.Our dental training system.The user is performing a drilling operation on the virtual teeth through haptic interface
ToM haptic device is used for haptic display.The dental training system is shown in Fig.1.
We proceed with a discussion on previous related work in Sect.2.Section 3discusses our haptic model and Sect.4describes haptic simulation techniques for dental training.Section Sect.5describes a dental workbench that provides an intuitive and immersive environment.We conclude in Sect.6.
2Related work
2.1V olume-based haptic rendering
The dental operations are implemented by detecting and removing carious lesions on the virtual tooth directly.In our system,the volumetric surface representation is used to simulate the diagnosis and cutting behavior through haptic interface.
In the haptic rendering for volumetric data,the force ?eld is typically computed directly from the volume data without conversion to the geometric model.A haptic ren-dering algorithm for volumetric data was introduced by Avila and Sobierajski [1].They used the gradient of the potential value to calculate the direction of the force and the amount of force was linearly proportional to the po-tential value.Kim et al.[9]used a hybrid surface represen-tation,which is a combination of the geometric model for visual rendering and the implicit surface for haptic display to take advantage of both data representations.
The volume-based haptic modeling system allows the user to intuitively add to and carve out material from a volumetric model through haptic interface.Avila and Sobierajski [1]introduced a haptic modeling system to di-rectly manipulate the volumetric model.Hua and Qin [8]suggested a physics-based volumetric modeling technique using dynamic spline-based implicit functions.Kim and Park [10]developed a haptic sculpting technique based on the volumetric implicit surface.
For dental simulation,we need to consider the hap-tic rendering techniques to simulate the tooth anatomy changes by the virtual dental tools.Typically,visual fre-quency (30Hz)is much slower than the haptic update rate (1KHz).This performance gap leads to force discontinu-ity if the force is computed directly on the physical surface of the tooth model being sculpted.In order to smooth the force rendering,[4,12]use the spring-based force estab-lished between the initial contact point on the surface and the current tool tip position.As a result,the user can move the tool tip freely and force computation is independent of the actual surface.Another method is “proxy blend-ing”[18],which smoothly interpolates the goal point from the old proxy point constrained on the old model to the new proxy point on the new surface.During the blending period,the user cannot adjust the blending speed and di-
92L.Kim,S.H.Park
rection,since the new surface should be de?ned when the blending starts.
2.2Dental training systems
DentSim manufactured by DenX Ltd.is a clinical simu-lator providing real-time tactile feedback with3D visual information.It is equipped with a patient mannequin,the typodont with a set of teeth,infrared camera and rotary dental instruments.As a student works with a tooth in the mannequin head,the computer generates a corresponding image of the tooth being prepared.In addition,DentSim can evaluate the student’s performance.The optical track-ing system requires time-consuming initialization and is costly.
Sae-Kee et al.,[19]suggested another dental training system,which comprises a force-torque measuring device and graphical animation.When the user performs any op-erations on the plastic tooth model mounted on the top of the sensor,the interactive simulation is graphically re-produced in the virtual environment using the measured force-torque data.Tooth tissue cutting simulation is imple-mented based on the volumetric tetrahedral model with the homogeneous tissue.However,there is no actual cutting process on the physical model,while drilling operations result in inconsistency between the virtual model and the actual one.
The Iowa Dental Surgical Simulator(IDSS)[22]was suggested to teach and evaluate the tactile and surgical skills relevant to the dental profession.However,the pro-totype only supports the detection of carious lesions(cav-ities)on the surface of teeth using the haptic probing sys-tem.The system provides visual information while physi-cal probing of the tooth surface.
Novint introduced the VRDTS(virtual reality dental training system)prototype[16,24].In this system,a stu-dent holds the PHANToM’s stylus(PHANToM is a haptic device from SensAble)and then uses it to probe a decayed virtual tooth,or to prepare the tooth using a virtual drill for cavity repair.This system allows one to restore the train-ing procedure and to quantify the student’s pro?ciency objectively.However,the system is at an early stage of development and does not provide detailed technical infor-mation.
3Volume-based haptic model
3.1Data representation
In our algorithm,a volumetric implicit surface is used for intuitive shape modi?cation without topological con-straints and haptic rendering.For visual display,it is con-verted into a geometric model that represents geometry ef-?ciently compared with volumetric representation.In the volumetric representation,only potential values close
to Fig.2.Data representation for surface modeling and haptic render-ing
the implicit surface(Fig.2)are involved in the computa-tion.
The potential values inside the close neighborhood of the surface range from?1to1according to the proxim-ity to the closet point on the surface.The values inside the surface are negative and positive outside.The values out of this narrow band are nothing to do with the surface modeling and haptic rendering.Therefore,to reduce the memory requirement,we use an octree-based data struc-ture,avoiding the representation of empty portions of the space.
A volumetric implicit surface representation is created from a geometric tooth model by Mauch’s fast closest point transform(CPT)algorithm[11].The algorithm?rst subdivides the3D space containing the geometric model in a regular3D grid.Then it computes the closest point to a surface and its distance at each grid point by solving the Eikonal equation using the method of characteristics.The volumetric representation is generated before the dental simulation.Figure3shows a volumetric teeth model con-verted from a geometric model by the CPT algorithm.In addition to potential?eld,tooth material properties such as stiffness and color are saved into the volumetric rep-resentation.Tooth anatomical tissue information can be obtained from micro CT data.Section4describes this in more
detail.
Fig.3.A volumetric teeth model is extracted from a geometric model by closest point transform.The potential?eld is sampled at each grid point(red points)in a volumetric representation
Haptic interaction and volume modeling techniques for realistic dental simulation
93
Fig.4a,b.Collision detection algorithm for volumetric model in previous systems.a Using a single point.b Using multiple points
3.2Collision detection
Most haptic rendering algorithms use only one point of the virtual tool for collision detection and force computation.However,the visual contact interaction between the vir-tual tool and the model surface may not correspond with the haptic feeling on complex models with sharp edges or small holes (see Fig.4(a)).Persik [15]uses multiple points distributed on the virtual tool to detect collision and cal-culate the force in a petrous bone surgery simulator (see Fig.4(b)).However,whenever the number and position of the points involved in force computation are changed,the system may lose stability and continuity in haptic sensa-tion with a buzzing sound.
In our system,we still use a single point of the vir-tual tool,but collision detection and force computation are performed on an offset surface rather than the im-plicit surface.The offset value from the implicit surface is determined by the radius of the bounding sphere of the tool.The center point of the bounding sphere becomes the single point.If the point collides with the offset sur-face,the system computes the force vector based on the offset surface (Fig.5).This approach can match both vi-sual and haptic sensation and provides stable force render-
ing.
Fig.5.Collision detection algorithm based on an offset surface.d is the offset value
3.3Force model
If the tool tip collides with the offset surface,the sys-tem computes the force vector.Equation Eq.1is used to compute the gradient,?f ,indicating the surface normal of an implicit surface.The force direction is computed
by interpolating the gradients of eight points of the cell containing the tool tip.The interpolation function leads the system to avoid force discontinuity.In order to deter-mine the amount of force,we ?rst ?nd the virtual contact point (VCP)on a surface,which is the intersection point between the surface and a ray along the computed force di-rection.The amount of force is proportional to the distance between the VCP and the tool tip.
?f =
d f dx ,d f dy ,d f dz ,(1)After th
e VCP is determined,a spring-damper model is used to compute the ?nal force vector in Eq.2,F =k (p c ?p t )?b (˙p c ?˙p t ),
(2)
where F is the force vector,p c and ˙p c are the coordi-nate and velocity of the VCP ,p t and ˙p t are the coordi-nate and the velocity of the tool tip,respectively,k is the stiffness,and b is the viscosity.The spring stiffness has a reasonably high value and the viscosity is to prevent os-cillations.
To simulate the friction on the virtual tooth model,we employed the physics-based model suggested by Hay-ward and Armstrong [7].The surface friction is imple-mented by simulating stick,creep and slip steps along with the movement of the tool.The creep step considers the transition to sliding motion with mass-spring behav-ior.
4Haptic simulation for dental training
4.1Multi-resolution mesh generation
Model modi?cation during sculpting is implemented by updating the local potential values around the sculpting tool,depending on the shape and size of the tool,opera-tion type (carving or adding)and the force applied by the user.The volumetric model is converted into the geometric model in every graphical frame for the visual rendering.Many volume sculpting systems [2,6]use a uniform poly-gonization method such as marching cubes,which suffers from a large triangle count and a resolution limited by a ?xed sampling rate.
In order to address these limitations,we employ the adaptive polygonization method suggested by Velho [23].This adaptive method consists of two steps.In the ini-tial polygonization step,a uniform space decomposition is used to create the initial mesh that serves as the basis for adaptive re?nement.In the second step,for adaptive poly-gonization,the mesh is re?ned adaptively at mid points of non-?at edges according to the surface curvature un-til the desired accuracy is achieved and then projected onto the implicit surface,since the mid point may not
94L.Kim,S.H.Park
be on the surface.The resulting mesh effectively repre-sents sharp edges with a smaller number of triangles and does not introduce cracks.The volumetric implicit sur-face and generated mesh are saved into another octree-based data structure to locally manage both data.The leaf level of the new octree structure is determined by the initial mesh resolution in the adaptive polygoniza-tion.
4.2Tooth material simulation
A tooth may be modeled by a multi-layer structure with different materials.The enamel layer is very hard and covers the outer surface of the tooth,dentin is the hard,porous,and yellow bone-like material that underlies the enamel and surrounds the entire nerve.Dental pulp is soft tissue containing blood vessels and nerves,and carious le-sion is the soft tissue in dark color.
In our system,tooth material properties are represented by stiffness and color information,which are saved into the volumetric representation.Different stiffness and col-ors correspond to different tissue.Tooth anatomical tissue information can be obtained through a Micro-CT scanner such as SkyScan TM [20].
The system allows dental students to feel the various visual,auditory,and haptic sensations during the dental procedures.Dental probing is to diagnose carious lesions (soft tissue in dark color)by a dental explorer.When the dental student is working on a decayed tooth using the dental drill,he can distinguish carious lesions from healthy part using different visual,auditory and haptic feedback.Color information provides visual feedback and stiffness in?uences the haptic resistance and physical con-tact sound.The stiffness value is calculated by averaging the stiffness values of the grid points within the local area intersected by the drill head.4.3Dental probing simulation
The dental student explores the tooth surface and detects the cavity by probing it with a pick.For advanced
lesions
Fig.6a,b.Dental probing to diagnose carious lesions.a Initial prob-ing before a preparation b Detecting if all caries has been removed after a
preparation
Fig.7a,b.Adaptive polygonization with the color complexity.a carious lesions on the tooth model.b multi-resolution mesh model
the sensation is similar to poking soft cheese.The system allows the student to diagnose carious lesions by a dental explorer both before and after the cavity preparation (see Fig.6).
To make various clinical cases,the system enables the user to directly assign stiffness and color values to the virtual tooth using a haptic interface (Fig.7(a)).Different colors indicates different material and anatomical parts.During mesh generation,color information in the volu-metric representation is used to assign per-vertex color.However,vertex color blends with neighbor vertices.To address this limitation,we use the adaptive polygoniza-tion approach in which the generated mesh is adapted to the color complexity,just as it is adapted to the surface complexity (Fig.7(b)).
4.4Cavity preparation using a drilling operation When the user performs the drilling operation,the sur-face of the virtual tooth should be modi?ed according to the shape of the applied tool.The system provides spherical and cylindrical tools as the dental drilling tool.Shape modi?cation is done by a CSG point-set operation between the drilling tool and 3D tooth model based on volumetric representation.The system updates potential values in a small local area around the tool (see Fig.8).The points that are not close to the surface may have
in-
Fig.8.Shape modi?cation of the virtual tooth model by a CSG point-set Boolean operation
Haptic interaction and volume modeling techniques for realistic dental simulation 95
herent errors,but this is acceptable for visual and haptic rendering.
The system typically could not update the graphical model at a rate suf?cient for haptic rendering (at least 1KHz).Due to this performance gap,most previous hap-tic sculpting systems were not able to directly simulate the graphical models being deformed.Instead,a virtual spring-based force [12]established between the current tool tip and the initial position is used to provide feedback.However,the user cannot directly feel the model surface being deformed and is not allowed to move the tool tip along the surface being deformed without retouching the surface.
To bridge the gap between the update rate of the visual and haptic rendering,we used the intermediate implicit surface suggested by Kim et al.[9].In this method,an in-termediate implicit surface smoothly transits between the discrete old model (current graphical model)and the tar-get models (next graphical model)in the haptic process.The target model is de?ned by a CSG (constructive solid geometry)point-set Boolean operation based on the old model and drilling tool.
The intermediate surface is animated from the old sur-face toward the target surface in the haptic loop.This means that the system updates the potential ?eld around the drilling tool to simulate the change of the intermedi-ate implicit surface.The amount of change depends on the force that the user applies,the local material stiff-ness,and the contact area between the virtual tooth and the drilling tool.Note that the force computation is im-plemented on the intermediate surface instead of graphical models in the haptic loop.In the next visual frame,the system updates the graphical model based on the current intermediate implicit surface at around 30Hz.If the user continues to drill,the system computes the next target sur-face by a CSG Boolean operation which is fast enough to avoid the impact in haptic sensation and then simulate the intermediate surface moving toward the target sur-face.
Figure 9shows a dental drilling procedure.The user re-moves carious lesions on the virtual teeth by using a cylin-drical drill
tool.
Fig.9a,b.Dental drilling simulation.a Removal of carious lesions on the top of the model using a cylindrical drill.b Removal of tooth tissue on the side of the tooth
model
Fig.10a,b.Filling the prepared cavities (see Fig.9)with amalgam (a )and removal of super?uous amalgam to match the original con-tour using a dental drill (b )
4.5Filling the prepared cavities
After preparing the teeth for cavity repair using dental drills,the system enables the dental users to ?ll the pre-pared cavity with amalgam (see Fig.10(a))and carve the amalgam to match the original tooth contour (see Fig.10(b)).
The procedure to ?ll the cavities is similar to the drilling procedure.The new target surface is computed by adding a CSG operation instead of a carving operation in the visual frame.The intermediate surface is animated from the current surface toward the target surface in the haptic loop.If the virtual tool touches the cavity,the local area around the tool swells up to ?ll the cavities.The newly added amalgam is represented in different color and stiffness to distinguish it from the original tissue both in terms of visual and haptic sensations.4.6Debris simulation
Dust and debris generated by the dental drill are simu-lated using a particle system.Each particle has a mass,position,velocity,and simple dynamic behavior.When the drilling operation on the virtual tooth is performed,par-ticles are randomly generated in the local area around the tool every visual frame.The mass of the particles decays as time passes.Figure 11shows a drilling simulation pro-
cess.
Fig.11.Dust simulation using a particle system
96L.Kim,S.H.Park
4.7Contact sound rendering
When the user applies the drilling tool on the virtual tooth model,physical contact sound enhances the realism sig-ni?cantly.In our system,parametric modeling techniques and vibration theory are used to simulate arti?cial sound.The AR (auto regressive)model [13]is used to simu-late a small drill sound.The sound rendering is executed in a different PC equipped with DSP and communicates with the main PC using TCP/IP .The main PC sends the sound PC messages regularly,containing information on power switch on/off,contact,and contact area.Drilling sound at idling and working is sampled and modeled as the AR model using the following equation:y (t )=a 1·y (t ?1)+···+a m ·y (t ?M )+e (t ),
(3)
where M is the model order,a 1,...,a m are AR coef?-cients,and e (t )is Gaussian noise.The AR(300)model is selected and the volume of drilling sound is adjusted ac-cording to contact area information to enhance reality.
5Dental workbench
In most of previous haptic systems,visual space and hap-tic space are not co-located,unlike in the real world.As a result,the user was not able to see or feel the virtual object in the same place.To provide more natural inter-face,Thomas et al.,[21]suggested the multi-modal visual workbench,which is designed to register three sensory modalities such as visual,haptic and auditory sensations.Similarly,we developed a dental workbench that allows dental students to see,hear,and touch the virtual tooth in the same virtual space.The dental workbench incor-porates stereo-visual display devices (stereo sync emitter and shutter glasses),a PHANToM haptic device,a Space-Mouse as a 6DOF position input device and an auditory display (see Fig.
12).
Fig.12.Dental workbench equipped with a visual,auditory,and haptic display
system
Fig.13.A user performing dental procedures on the dental work-bench
The user sees the images on a mirror re?ecting the monitor mounted on the dental workbench (see Fig.13).For the stereoscopic view,we use GL_STEREO quad buffering on a stereo-ready graphics card such as NVidia FX 3000.A common approach to make the stereoscopic view is to rotate the cameras and point at the center of the object (see Fig.14(a)).However,once the cam-eras are rotated,their axes of projection become dif-ferent introducing vertical parallax,which requires the stereo “point of focus”to be constantly adjusted when the viewer is moving through the scene.To address this limitation,the system skews each camera’s view frustum so that view vectors for each eye are parallel (see Fig.
14(b)).
Fig.14a,b.Stereo rendering.a Common approach b Asymmetric view frustum
6Conclusion
We introduced haptic dental simulation techniques based on a volumetric representation containing implicit surface and material properties.The system allows dental stu-dents to learn dental procedures including detecting the decayed tooth,removing carious lesions by a drilling operation,and ?lling the prepared cavities with amal-gam.For haptic rendering,collision detection and smooth force computation are implemented based on the offset surface and tooth materials represented by volume stiff-
Haptic interaction and volume modeling techniques for realistic dental simulation97
ness and color information are simulated.The volumetric tooth model can be modi?ed during the drilling(carv-ing)and?lling(adding)operations and then converted into a geometric model using an adaptive polygonization method every visual frame.To enhance the realism,the system simulates debris and physical contact sound be-tween the virtual tooth model and the drilling tool.Fi-nally,we introduced a dental workbench to allow the users to see,hear,and feel the virtual model in the same space.
We plan to enhance the dental training system by adding various dental tools,undo/redo functions,evalu-ation of student performance,and recoding the training procedure performed by dental students.In addition,we will collaborate with dental schools to evaluate our system by dental instructors and students.
References
1.Avila,R.S.,Sobierajski,L.M.:A haptic
interaction method for volume
visualization.IEEE Visualization
Proceedings,pp.197–204(1996)
2.Baerentzen,J.A.:Octree-based volume
sculpting.IEEE Visualization Proceedings, pp.9–12(1998)
3.Bloomenthal,J.:Introduction to Implicit
surface.Morgan Kaufmann Publishers,San Francisco,CA(1997)
4.Foskey,M.,Otaduy,M.A.,Lin,M.C.:
ArtNova:Touch-enabled3D model design.
IEEE Virtual Reality Proceedings,
pp.119–126(2002)
5.SensAble Technologies Inc.:Freeform
modeling system,
https://www.360docs.net/doc/fd13385007.html,/freeform(1999) 6.Galyean,T.A.,Hughes,J.F.:Sculpting:an
interactive volumetric modeling technique.
ACM SIGGRAPH Proceedings,
pp.267–274(1991)
7.Hayward,V.,Armstrong,B.:A new
computational model of friction applied to haptic rendering.Experimental Robotics,
Lect.Notes Comput.Sci.250,404–412
(2000)
8.Hua,J.,Qin,H.:Haptic sculpting of
volumetric implicit functions.Paci?c
Conference on Computer Graphics and
Applications Proceedings,pp.254–264,
Tokyo(2001)
9.Kim,L.,Sukhatme,G.S.,Desbrun,M.:
A haptic rendering technique based on
hybrid surface represenation.IEEE Comput Graph Appl,Special Issue–Haptic
Rendering Beyond Visual Computing
24(2),66–75(2004)
10.Kim,L.,Park,S.:A Haptic Sculpting
Technique Based on V olumetric
Representation.Lecture Notes in Computer
Science,V ol.3179.Springer(2004)
11.Mauch,S.:A fast alogorithm for computing
the closest point and distance transform.
Technical Report at
https://www.360docs.net/doc/fd13385007.html,/～sean/(2000)
12.McDonnell,K.T.,Qin,H.,Wlodarczyk,
R.A.:Virtual clay:a real-time sculpting
system with haptic toolkits.ACM
Symposium on Interactive3D Techniques,
pp.179–190(2001)
13.Pandit,S.M.,Wu,M.S.:Time Series and
System Analysis with Applications.John
Wiley,New York(1983)
14.Perry,R.N.,Frisken,S.F.:Kizamu:
A system for sculpting digital characters.
ACM SIGGRAPH,pp.47–56(2001)
15.Petersik,A.,P?esser,B.,Tiede,U.,Hohne,
K.,Leuwer,R.:Realistic Haptic Interaction
in V olume Sculpting for Surgery
Simulation.Lecture Notes in Computer
Science,V ol.2673,pp.194–202.Springer
(2003)
16.Ranta,J.F.,Avile,W.A.:The virtual reality
dental tranining system–simulating dental
procedures for the purpose of training
dental students using haptics.Fourth
Phantom User Groups Workshop
Proceedings,pp.73–77(1999)
17.Ruspini,D.C.,Kolarov,K.,Khatib,O.:The
haptic display of complex graphical
environment.ACM SIGGRAPH
Proceedings,V ol.1,pp.295–301(1997)
18.Ruspini,D.C.,Khatib,O.:Dynamic models
for haptic rendering systems.Advances in
Robot Kinematics Proceedings,
pp.523–532(1998)
19.Sae-Kee,B.,Riener,R.,Frey,M.,Proll,T.,
Burgkart,R.:Phantom-based interactive
simulation system for dental treatment
training.Annual Medicine Meet Virtual
Reality(2004)
20.SkyScan:microtomography,
nanotomography,3D x-ray microschopy,
http://www.skyscan.be/
21.Wiegand,E.T.,Schloerb,W.D.,Sachtler,
W.L.,Virtual Workbench:Near-?eld virtual
environment system with applications.
Presence8(5),492–519(1999)
22.Thomas,G.,Johnson,L.,Dow,S.,
Standford,C.:The design and testing of
a force feedback dental https://www.360docs.net/doc/fd13385007.html,put.
Meth.Prog.Biomed.64,53–64(2000)
23.Velho,L.,Figureiredo,L.H.D.,Gomes,J.:
A uni?ed approach for hierarchical adaptive
tesselation of surfaces.ACM Trans.Graph.
18(4),329–360(1999)
24.Virtual Reality Dental Training System
(VRDTS),
https://www.360docs.net/doc/fd13385007.html,/VRDTS.htm
25.Zilles,C.,Salisbury,J.K.:
A constraint-based God-object method for
haptic display.Haptic Interfaces for Virtual
Environment and Teleoperator Systems
Proceedings,pp.146–150(1994)
98L.Kim,S.H.
Park
L AEHYUN K IM is a research scientist in the
Systems Technology Division at Korea Institute
of Science Technology.His research interests
include haptics,computer graphics,and virtual
reality.Kim received a BS degree from Hanyang
University in Material Engineering and an MS
degree from Yonsei University in Computer
Science,and a PhD degree in Computer Science
from the University of Southern Califronia.
S E H YUNG P ARK is a principal researcher in the
Systems Technology Division at Korea Institute
of Science Technology.His research interests
include the study of geometric modeling,human
computer interface,reverse engineering and
NC programming.Park received a BS degree
from Seoul National University in Mechanical
Design and Production Engineering and an MS
degree from Cornell Unversity in Mechanical
Engineering,and a PhD degree in Mechnical
Engineering from Korea Advanced Institute of
Science and Technology.