Presentation is loading. Please wait.

Presentation is loading. Please wait.

Fast Texture Synthesis using Tree-structured Vector Quantization Li-Yi Wei Marc Levoy Computer Graphics Group Stanford University.

Similar presentations


Presentation on theme: "Fast Texture Synthesis using Tree-structured Vector Quantization Li-Yi Wei Marc Levoy Computer Graphics Group Stanford University."— Presentation transcript:

1 Fast Texture Synthesis using Tree-structured Vector Quantization Li-Yi Wei Marc Levoy Computer Graphics Group Stanford University

2 Introduction Texture Synthesis Input Result

3 Desirable Properties Result looks like the input Efficient General Easy to use Extensible

4 Previous Work Procedural Synthesis –Perlin 85, Witkin 91, Worley 96 Statistical Feature Matching –Heeger 95, De Bonet 97, Simoncelli 98 Markov Random Fields –Popat 93, Efros 99

5 Outline Basic algorithm Multi-resolution algorithm Acceleration Applications

6 Texture Model Textures are –local –stationary Model textures by –local spatial neighborhoods

7 Basic Algorithm Exhaustively search neighborhoods

8 Neighborhood Use causal neighborhoods CausalNon-causal Input Noise

9 Neighborhood Neighborhood size determines the quality & cost 333355557777 9999 11  1141  41 423 s528 s 739 s 1020 s1445 s 24350 s

10 Multi-resolution Pyramid High resolutionLow resolution

11 Multi-resolution Algorithm

12 Benefit Better image quality & faster computation 1 level 5  5 3 levels 5  5 1 level 11  11

13 Results Random Oriented RegularSemi-regular

14 Failures Non-planar structures Global information

15 Comparison Heeger 95De Bonet 97Efros 99Our method Input 1941 secs 503 secs 12 secs

16 Acceleration Computation bottleneck: neighborhood search

17 Nearest Point Search Treat neighborhoods as high dimensional points 1 2 3 4 5 6 7 8 9 10 11 12 Neighborhood 1 2 3 4 5 6 7 8 9 10 11 12 High dimensional point/vector

18 Acceleration Nearest point search in high dimensions –[Nene 97] Cluster-based model for textures –[Popat 93] Tree-structured Vector Quantization –[Gersho 92]

19 Tree-structured Vector Quantization

20 Timing Time complexity : O(log N) instead of O(N) –2 orders of magnitude speedup for non-trivial images 1941 secs503 secs12 secs Efros 99Full searchingTSVQ

21 Results: Brodatz Textures InputExhaustive: 360 secsTSVQ: 7.5 secs D103 D20

22 Application 1: Constrained Synthesis ?

23 Possible Solution Multi-resolution blending [Burt & Adelson 83] –produce visible boundaries

24 Possible Solution Original raster-scan algorithm –discontinuities at right and bottom boundaries

25 Possible Solution Adaptive neighborhoods [Efros 99] –Hard to accelerate

26 Modifications Need to use a single symmetric neighborhood 2 pass algorithm with extrapolation Spiral order synthesis

27 Result

28 Result Extrapolation ? ? ? ?

29 Result Image editing by texture replacement

30 Application 2: Temporal Texture Indeterminate motions both in space and time –fire, smoke, ocean waves How to synthesize? –extend our 2D algorithm to 3D

31 Temporal Texture FireSmokeWaves Input Result

32 Future Work More general “textures” –light fields, solid textures –motion signals –displacement maps Real time texture synthesis

33 Acknowledgment Kris Popat Alyosha Efros Stanford Graphics Group Intel, Interval, Sony More information http://graphics.stanford.edu/projects/texture/


Download ppt "Fast Texture Synthesis using Tree-structured Vector Quantization Li-Yi Wei Marc Levoy Computer Graphics Group Stanford University."

Similar presentations


Ads by Google