existing context models
play

Existing Context models Torralba,*Sinha*(2001)* - PowerPoint PPT Presentation

PanoContext A Whole-room 3D Context Model for Panoramic Scene Understanding by Yinda Zhang, Shuran Song, Ping Tan, Jianxiong Xiao Presented by: William Xie Existing Context models Torralba,*Sinha*(2001)*


  1. PanoContext A Whole-room 3D Context Model 
 
 for Panoramic Scene Understanding by Yinda Zhang, Shuran Song, Ping Tan, Jianxiong Xiao Presented by: William Xie

  2. Existing Context models Torralba,*Sinha*(2001)* Carbone(o,*de*Freitas*&*Barnard*(2004)* Torralba*Murphy*Freeman*(2004)* Rabinovich*et*al*(2007)* Fink*&*Perona*(2003)* Heitz*and*Koller*(2008)* Kumar,*Hebert*(2005)* Sudderth,*Torralba,* Wilsky,*Freeman*(2005)** Hoiem,*Efros,*Hebert*(2005)* Desai,*Ramanan,*and*Fowlkes*(2009)* DPM$on$PASCAL$VOC$[Felzenszwalb$et$al.]$ Improvement on PASCAL <1.5% Slide credit: Zhang et al.

  3. What is this object? Slide credit: Zhang et al.

  4. What is this object? Slide credit: Zhang et al.

  5. What is this object? Slide credit: Zhang et al.

  6. What is this object? Slide credit: Zhang et al.

  7. What is this object? Slide credit: Zhang et al.

  8. What is this object? Slide credit: Zhang et al.

  9. Why didn’ context help?

  10. Why didn’ context help? Perhaps we are not using the right data

  11. PASCAL VOC • On average: 1.5 object classes and 2.7 object instances per image • Average camera field of view: 40° - 60° horizontal

  12. Human Vision • 180° horizontal field of view • Ability to see depth • Ability to change viewpoint

  13. Remedy

  14. PanoContext Slide credit: Zhang et al.

  15. PanoContext Input: Panorama Slide credit: Zhang et al.

  16. PanoContext door painting mirror painting tv window desk bed sofa chair nightstand Input: Panorama Output: 2D projected result painting mirror ��� window ��� door desk ��� tv �� � painting ��� bed ���� ��� ���� ��� ! ���� sofa ���� ��� nightstand ���� ��� ��� � ��� �� ��� � ��� ��� Output: 3D model Slide credit: Zhang et al.

  17. PanoContext door painting mirror painting tv window desk bed sofa chair nightstand Input: Panorama Output: 2D projected result painting mirror ��� window ��� door desk ��� tv �� � painting ��� bed ���� ��� ���� ��� ! ���� sofa ���� ��� nightstand ���� ��� ��� � ��� �� ��� � ��� ��� Output: 3D room exploration Output: 3D model Slide credit: Zhang et al.

  18. Pipeline

  19. Pipeline Krizhevsky, Alex, et al. "Imagenet classification with deep convolutional neural networks." NIPS. 2012.

  20. Pipeline • Vanishing point estimation for panoramas • Room layout hypothesis generation • 3D object hypotheses generation • Whole-room scene hypotheses generation • Data-driven holistic ranking

  21. Pipeline • Vanishing point estimation for panoramas • Room layout hypothesis generation • 3D object hypotheses generation • Whole-room scene hypotheses generation • Data-driven holistic ranking

  22. Pipeline • Vanishing point estimation for panoramas • Room layout hypothesis generation … … • 3D object hypotheses generation • Whole-room scene hypotheses generation • Data-driven holistic ranking

  23. Pipeline • Vanishing point estimation for panoramas • Room layout hypothesis generation … … • 3D object hypotheses generation • Whole-room scene hypotheses generation ✓ ! • Data-driven holistic ranking

  24. Generate a pool of hypotheses Whole Input Room Room Object Slide credit: Zhang et al.

  25. Generate a pool of hypotheses Whole Input Room Room Object Slide credit: Zhang et al.

  26. Room layout hypothesis Slide credit: Zhang et al.

  27. Room layout hypothesis Line segments detection Algorithm Slide credit: Zhang et al.

  28. Room layout hypothesis Hough transform for vanishing point Slide credit: Zhang et al.

  29. Room layout hypothesis Hough transform for vanishing point Classify a vanishing direction for each line Slide credit: Zhang et al.

  30. Source: Wikipedia, Emaze

  31. Room layout hypothesis Sample 5 line segments to generate a room layout

  32. Room layout hypothesis Sample 5 line segments to generate a room layout

  33. Room layout hypothesis Sample 5 line segments to generate a room layout Slide credit: Zhang et al.

  34. Room layout hypothesis Sample 5 line segments to generate a room layout Slide credit: Zhang et al.

  35. Room layout hypothesis Sample 5 line segments to generate a room layout Slide credit: Zhang et al.

  36. Room layout hypothesis Sample 5 line segments to generate a room layout Slide credit: Zhang et al.

  37. Room layout hypothesis Pixel-wise surface direction estimation Slide credit: Zhang et al.

  38. Room layout hypothesis Line segments Slide credit: Zhang et al.

  39. Room layout hypothesis Line segments Slide credit: Zhang et al.

  40. Room layout hypothesis Line segments Surface normal estimation Slide credit: Zhang et al.

  41. Room layout hypothesis Line segments Surface normal estimation Slide credit: Zhang et al.

  42. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 Slide credit: Zhang et al.

  43. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 Slide credit: Zhang et al.

  44. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 0.711 Slide credit: Zhang et al.

  45. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 0.711 Slide credit: Zhang et al.

  46. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 0.711 0.504 Slide credit: Zhang et al.

  47. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 0.711 0.504 Slide credit: Zhang et al.

  48. Room layout hypothesis Line segments Surface normal estimation Consistency Score: 0.770 0.711 Slide credit: Zhang et al.

  49. Room layout hypothesis Line segments Top 50 only Surface normal estimation Consistency Score: 0.770 0.711 Slide credit: Zhang et al.

  50. Generate a pool of hypotheses Whole Input Room Room Object Slide credit: Zhang et al.

  51. Cuboid detection Input: a single-view panorama Slide credit: Zhang et al.

  52. Cuboid detection Input: a single-view panorama Fitted cuboids Slide credit: Zhang et al.

  53. Cuboid detection DPM-esque 6 rays and Largest IoU Selective search 3 vanishing points with the segment Slide credit: Zhang et al.

  54. Semantic classification Random Object Features forest categories • Size bed • Aspect ratio & Area desk • Distance to walls sofa … chair Slide credit: Zhang et al.

  55. Semantic classification Random Object Features forest categories • Size bed • Aspect ratio & Area desk • Distance to walls sofa … chair 70% Accuracy Slide credit: Zhang et al.

  56. Semantic classification bed Slide credit: Zhang et al.

  57. Semantic classification nightstand Slide credit: Zhang et al.

  58. Semantic classification painting Slide credit: Zhang et al.

  59. Pairwise constraint Slide credit: Zhang et al.

  60. Generate a pool of hypotheses Whole Input Room Room Object Slide credit: Zhang et al.

  61. Data-driven sampling Randomly sample a room layout With P(layout) ∝ normal consistency score Slide credit: Zhang et al.

  62. Data-driven sampling Randomly sample a room layout With P(layout) ∝ normal consistency score Slide credit: Zhang et al.

  63. Data-driven sampling Slide credit: Zhang et al.

  64. Data-driven sampling Decide number of object based on prior distribu<on: paintin 2 g bed 1 desk 1 nightst 1 and mirror 1 sofa 1 tv 1 window 1 Slide credit: Zhang et al.

  65. Data-driven sampling Decide number of object Decide object sampling sequence based on prior distribu<on: based on bo?om up scores: paintin bed 2 g nightstand bed 1 painting desk 1 desk nightst 1 window and mirror 1 painting sofa 1 tv tv 1 sofa window 1 mirror Slide credit: Zhang et al.

  66. Data-driven sampling Sample a bed in empty room first… Confidence Low High 10 20 30 40 50 60 70 Bottom-up score as bed Slide credit: Zhang et al.

  67. Data-driven sampling Sample a bed in empty room first… Randomly select one according to bottom up priority rectangle detection score, semantic classifier score Slide credit: Zhang et al.

  68. Data-driven sampling Then, sample a nightstand given a bed Randomly select one according to the bottom up + pair-wise priority mean distance to the K nearest neighbors Slide credit: Zhang et al.

  69. Pairwise constraint Slide credit: Zhang et al.

  70. Data-driven sampling Keep on sampling until finishing the list… painting bed nightstand List: bed, nightstand, painting, desk, window, painting, TV, sofa, mirror Slide credit: Zhang et al.

  71. Data-driven sampling Keep on sampling until finishing the list… painting desk bed nightstand List: bed, nightstand, painting, desk, window, painting, TV, sofa, mirror Slide credit: Zhang et al.

Recommend


More recommend