This preview shows pages 1–10. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 1 1 What we will cover • Contour Tracking • Surface Rendering • Direct Volume Rendering • Isosurface Rendering • Optimizing DVR • PreIntegrated DVR 2 When we stop resampling along the ray? What voxels we can skip during rendering? How to improve resampling process? How to find the nearest object boundary? How to improve image quality? Optimizing DVR 2 3 Early ray termination Terminate resampling when the accumulated opacity reaches the threshold value. Image plane pixel Volume data 4 Using data coherency • Good for empty space skipping • Octree 1 2 3 4 5 6 • • • • • • root 2 1 3 4 5 6 7 8 3 5 Using data coherency • Kd tree – Recursively subdivide the volume along x,y, and zaxis aligned planes. • Run length encoding – encoded by first location of run and length of run 6 • Traverse a discrete representation of the ray • 3D line scanconversion or voxelization algorithm • Three types of connected paths Discrete ray casting 4 7 Template based rendering • Accelerating ray casting by minimizing resampling time – use inter ray coherence • Need a different template per individual displacement of a ray in a image pixel • No interactive speed because of image order processing template Image plane ray 8 Polygon assisted ray casting • Efficient empty space skipping • But the more twisted the object is, the more polygons are needed 5 9 Distance map Good for empty space skipping Good for preclassified volume Each voxel contains the shortest distance to the boundary for visualization Need extensive preprocessing time When the distance is too big, set the maximum value for reducing memory space for distance. 10 Distance map – example Traversal time with Distance map : 2 times Without Distance map : 5 times 6 11 P. Lacroute and Marc Levoy[94] Image and object space method Very fast s/w based algorithm Need preprocessing step for encoding Preclassification with opacityweighted colors are common Shearwarp algorithm 12 Shearwarp algorithm 1. Transform the volume data to sheared object space by translating and resampling each slice (also scaling for perspective transformation) warp shear Project warping 7 13 Shear warp rendering 2. Composite the resampled slices together in frontto back order using the “over” operator 3. Transform the intermediate image to image space by warping it according to warp view M S P M view warp M P S M 1 1 warp M 14 Shear warp rendering 8 15 Shear warp rendering • Runlength encoding 16 Shearwarp rendering Pros Fast : image & object space algorithm Simple perspective projections possible hardware acceleration possible Cons: Not good image quality bilinear interpolation & warping distortion voxel/pixel=1: problems for zooming Require three sets of encoded volume 9 17 Shearwarp vs. Raytracing nearest neighbor interpolation linear interpolation...
View
Full
Document
 Spring '09
 Shinyeonggil
 Computer Graphics

Click to edit the document details