1 / 28

期末 DEMO line segment 的 GPU 實現

期末 DEMO line segment 的 GPU 實現. 2012/08/14 指導教授:詹寶珠教授 報告者:陳正旻. Outline. 演算法流程介紹 ( 附實作結果 ) 時間比較和加速倍數 實作影片 總結. Flow chart. Algorithm. Line segment construction Two rules Color difference : Spatial distance :. (TIME)Line segment construction. Flow chart. Algorithm. Left. Right.

naida
Télécharger la présentation

期末 DEMO line segment 的 GPU 實現

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. 期末DEMO line segment的GPU實現 2012/08/14 指導教授:詹寶珠教授 報告者:陳正旻 1

  2. Outline • 演算法流程介紹(附實作結果) • 時間比較和加速倍數 • 實作影片 • 總結 2

  3. Flow chart 3

  4. Algorithm • Line segment construction • Two rules • Color difference : • Spatial distance : 4

  5. (TIME)Line segment construction 5

  6. Flow chart 6

  7. Algorithm Left Right d=1 R. Zabih and J. Woodfill, “Non-parametric local transforms for computing visual correspondence,” in Proc. ECCV, 1994, pp. 151–158. String 2 • Initial Disparity Computation • Matching cost : • CAD : • CCENSUS : Hamming distance of the two strings that stand for p and pd 7

  8. (TIME)Initial Disparity Computation& Matching cost 8

  9. Algorithm Left Right d=1 Cost aggregation Winner-take-all 9

  10. (TIME)Cost aggregation& Winner-take-all 10

  11. 初步結果 左圖的Depth 右圖的Depth 11

  12. Flow chart 12

  13. Algorithm Seed Pixel Detection Two rules 13

  14. (TIME)Seed Pixel Detection 14

  15. Flow chart 15

  16. Algorithm • Scanline Propagation • Aa , if p is in occluded regions. • d • Ss , if p is near depth discontinuities. • a • is replaced by the linearly interpolated results of and in all other conditions. 16

  17. (TIME)Scanline Propagation 17

  18. Scanline Propagation之後的結果 18

  19. Flow chart 19

  20. Algorithm • Disparity Refinement • Vertical voting : We collect 𝑁 disparity votes in a vertical line segment that starts at pixel and ends at pixel ,if the color difference is smaller than the threshold value 𝜏 , the votes on disparity increase by 1. • c 20

  21. (TIME)Vertical voting 21

  22. Vertical voting之後的結果 22

  23. 再加入BilateralFilter之後的最終結果 23

  24. GPU運算時間 CPU運算時間 24

  25. 時間比較和加速倍數 25

  26. 實作影片 26

  27. 總結 GPU加速持續改進。 再加入各種不同方法使結果更好。 27

  28. Thank for your attention! 28

More Related