VIRTUAL ARTHROSCOPIC KNEE SURGERY TRANING SYSTEM Yang Xiaosong The Chinese University of Hong Kong Tsinghua University.

VIRTUAL ARTHROSCOPIC KNEE SURGERY TRANING SYSTEM

Yang XiaosongThe Chinese University of Hong Kong

Tsinghua University

VIRTUAL ARTHROSCOPIC KNEE SURGERY TRANING SYSTEM

A joint project between

the Chinese University of Hong Kongthe Chinese University of Hong Kong

Tsinghua University,Tsinghua University,

sponsored by

The National Natural Science Foundation of China RGC of Hong Kong

Minimally Invasive Microsurgical Technique

Less trauma Reduced pain Quicker convalescence

Restrictions of Arthroscopy

Restricted vision Poor hand-eye coordination Limited mobility of surgical instruments

Surgical Skill Training

Animals Cadavers Virtual reality based simulation systems

Arthroscopy Surgery

Virtual Arthroscopic Knee Surgery Training System

Modeling using data from Visible Human Project Simulation of the deformation of soft tissue with

topological change by FEA User interaction

– Force feedback

Hardware System Architecture

Central Computer(PIV 1.5G, Nivdia Geforce 3,

Windows 2000)

Input Device Display Screen

Software System Architecture

3D Segmentation3D Segmentation

CT, MRI Volume Data

Segmented Volume Data Geometry Modeling

Geometry Modeling

Preprocess Stage

Real-time simulation of non-linear deformation with cutting

Force feedback calculation of soft tissues

Real-time simulation of non-linear deformation with cutting

Force feedback calculation of soft tissues

Force Feedback DeviceForce Feedback Device

Physical Attributes

Set Force

Collision DetectionCollision Detection

Manipulation of Operation Facilities Contacted

Realistic RenderingView from outsideArthroscopy

Realistic RenderingView from outsideArthroscopy

Surface mesh

Simplify & SmoothSimplify & Smooth

Local Remesh in Operation Area

Local Remesh in Operation Area

On the FLY StageSurface and Tetrahedral mesh

3D Tetrahedral mesh

Surface

mesh

Mesh Generation of Human Organs

Segmentation Surface boundary meshes creation Tetrahedral mesh generation Mesh smoothing

Collision Detection

Prevent the arthroscope and operation facility from entering a solid object

Get the initial intersection point for cutting simulation

Collision detection for deformable objects, different from that of rigid objects

AABB tree

Simulation of Soft Tissue Deformation With Flexible Cutting

Physically reality Real-time interaction

Hybrid Finite Element MethodHybrid Finite Element Method

Hybrid FEM

Non-linear deformation and topology changing model in operating region (Region 1).

– The local small region, fast to compute Linear deformation and topology constant model in non-operating region

(Region 2)– The remaining large region, accelerated by pre-processing

2-Dimension Sample

Cutting of a single element

Normal Cases Degeneration cases

3-Dimension Example

A simplified model of thigh

Tetrahedral meshes simplification

Input Device

Four DOFs for arthroscope and instruments– Pitch– Yaw– Insertion– Rotation

Force feedback– Three DC motors attached for the first three DOF– The fourth DOF need no force feedback

Input Device Picture

System Interface

Sample

Work to do

More effective interactive 3-D segmentation system Realistic Rendering Simulation of complicated operation facilities

Tetrahedral Mesh Generation of Human Organs on Segmented Volume

Tetrahedralization Algorithm on Segmented Volume

Voxel-Split tetrahedralization

3D conforming Delaunay tetrahedralization algorithm

Feature point based tetrahedralization

Voxel-Split tetrahedralization

Voxel-Split tetrahedralization

Voxel-Split tetrahedralization

Voxel-Split tetrahedralization

Voxel-Split tetrahedralization

Simplification of Segmentation Volume

Voxel-Split tetrahedralization

Global Simplification of Segmentation Volume

– Boundary voxel decomposition

– The order of voxel merge

3D conforming Delaunay tetrahedralization algorithm

Tissue boundaryextraction

3D conforming Delaunay tetrahedralization algorithm

Feature point based tetrahedralization

Accurate.

Small scale.

Well-shaped.

Feature point based tetrahedralization

Placement of the mesh vertices Delaunay Triangulation Restore the tissue boundary and set element’s tis

sue type

Feature point based tetrahedralization

Point Displacement – feature point, steiner point and structured mesh point

Feature point based tetrahedralization

Feature Point

Feature point based tetrahedralization

Feature Point

1. Gradient computation of the mid point of each voxel edge2. Compare of the gradient in the local neighbors3. Error bounded simplification of feature point

Feature point based tetrahedralization

Steiner point displacement

n

kn

k

n

kjj

kjj

k

zyxd

zyxd

DzyxD1

1

),,(

),,(

),,(

Feature point based tetrahedralization

Feature point based tetrahedralization

Cross Tissue Boundary Detection

• Criterion for crossing boundary

• Boundary Points (BP)

• Voxel Points (VP)

• Edge: VP-VP BP-VP BP-BP

Feature point based tetrahedralization

Remesh to restore the tissue boundary

Flip 23

Flip 32

c

f

e

d a b

c

f

e

d a b

Flip4Diagonal

• No=3, flip32 to delete the crossing edge.• No=4, Flip4Diagonal to swap the diagonal crossing edge.• No>4

Feature point based tetrahedralization

Remesh to restore the tissue boundary

Feature point based tetrahedralization

Volume 297 x 341 x 180= 18,229,860

Tetehedral Mesh 94,953 nodes, 490,409 elements

Thanks