registration for mobile hdr
play

REGISTRATION FOR MOBILE HDR PHOTOGRAPHY Orazio Gallo, 04/06/2016 - PowerPoint PPT Presentation

April 4-7, 2016 | Silicon Valley LOCALLY NON-RIGID REGISTRATION FOR MOBILE HDR PHOTOGRAPHY Orazio Gallo, 04/06/2016 (work with Alejandro Troccoli, Jun Hu, Kari Pulli, and Jan Kautz) WHAT IS HIGH -DYNAMIC- RANGE? Kinda depends on whom you


  1. April 4-7, 2016 | Silicon Valley LOCALLY NON-RIGID REGISTRATION FOR MOBILE HDR PHOTOGRAPHY Orazio Gallo, 04/06/2016 (work with Alejandro Troccoli, Jun Hu, Kari Pulli, and Jan Kautz)

  2. WHAT IS “HIGH -DYNAMIC- RANGE”? Kinda depends on whom you ask… Displays Videogames Photography Photometry … 2

  3. WHAT IS HDR? http://www.flickr.com/photos/lprowler/5704117093/ 3

  4. WHAT IS HDR? 4

  5. WHAT IS HDR? 6

  6. WHY DO WE NEED REGISTRATION? 7

  7. WHY DO WE NEED REGISTRATION? 8

  8. WHY DO WE NEED REGISTRATION? 9

  9. WHY DO WE NEED REGISTRATION? 10

  10. RELATED WORK 11

  11. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 12

  12. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 13

  13. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 14

  14. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 15

  15. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 16

  16. FULLY NON-RIGID REGISTRATION Hu, Gallo, Pulli , Sun, “ HDR Deghosting: How to Deal with Saturation? ” IEEE CVPR 2013. 17

  17. RELATED WORK Speed Rigid registration Milliseconds (Ward ‘03, Tomaszewska and Mantiuk ‘07, …) Accelerated patch- based (Bao et al. ‘14) Seconds Rejection methods ( Gallo ‘09, Zhang and Cham ‘12, Oh et al. ‘15 …) Flow-based methods (Zimmer et al. ‘11, Zhang and Cham ‘12,…) Fully non-rigid registration Minutes (Hu et al. ‘12, Sen et al. ‘12, Hu et al. ‘13) Parallax Motion artifacts Little to no Quality and or artifacts 19 motion artifacts reduced dynamic range

  18. WHAT’S THE CATCH? Original SIFT+warp Ours 25

  19. A SPARSE-TO-DENSE APPROACH Compute the flow at sparse locations, Propagate the flow in an edge-aware fashion, and Merge the images in an error-tolerant way. 28

  20. METHOD 29

  21. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion 30

  22. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference Source 31

  23. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference Source 34

  24. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference Source 35

  25. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference 39

  26. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Source 40

  27. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference 41

  28. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Source 42

  29. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference 43

  30. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Source 44

  31. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference’s Luma Sparse flow 57

  32. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion sparse samples latent signal 59

  33. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion filtered sparse samples 60

  34. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion sparse samples normalization map 61

  35. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion filtered sparse samples reconstructed signal filtered normalization map 62

  36. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Luminance Pixel 67

  37. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Luminance Pixel 68

  38. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Luminance Pixel Gastal and Oliveira, “ Domain transform for edge-aware image and video processing ” ( SIGGRAPH '11) 69

  39. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion 1 u v 1 Reference Luma, L Sparse flow, f Normalization map N 70

  40. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference 71

  41. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Source 72

  42. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Reference 73

  43. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Warped source 74

  44. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion Warped source 75

  45. Corners and Sparse-to- Error-tolerant Filter matches matches dense warp fusion 76

  46. PERFORMANCE 77

  47. Execution time for a pair of 5MP images 78

  48. Related work Speed Rigid registration Milliseconds (Ward ‘03, Tomaszewska and Mantiuk ‘ 07, …) Accelerated patch-based (Bao et al. ‘14) Seconds Rejection methods (Gallo ‘09, Zhang and Cham ‘ 12, Oh et al. ‘ 15 …) Flow-based methods (Zimmer et al. ‘11, Zhang and Cham ‘ 12,…) Fully non-rigid registration Minutes (Hu et al. ‘12, Sen et al. ‘12, Hu et al. ‘ 13) Parallax Motion artifacts Quality No artifacts and or 79 motion artifacts reduced dynamic range

  49. Related work Speed Milliseconds Accelerated patch-based (Bao et al. ‘14) Seconds (VGA resolution) Fully non-rigid registration Minutes (Hu et al. ‘12, Sen et al. ‘12, Hu et al. ‘ 13) Parallax Motion artifacts Quality No artifacts and or 80 motion artifacts reduced dynamic range

  50. Related work Speed Milliseconds Accelerated patch-based (Bao et al. ‘14) Seconds (5MP resolution) Fully non-rigid registration Minutes (Hu et al. ‘12, Sen et al. ‘12, Hu et al. ‘ 13) Parallax Motion artifacts Quality No artifacts and or 81 motion artifacts reduced dynamic range

  51. What about visually? Ours Bao et al.’s (2.5MP) 82

  52. TO SUM UP… Contributions A fast registration algorithm >11x faster than the fastest published method We propose to use a sparse-to-dense approach CUDA-based sparse-to-dense propagation CUDA-based robust image fusion 83

  53. MORE RESULTS Reference 85

  54. MORE RESULTS Source 86

  55. MORE RESULTS Reference 87

  56. MORE RESULTS Warped source 88

  57. MORE RESULTS HDR 89

  58. MORE RESULTS Naïve fusion Our result 90

  59. MORE RESULTS Reference 91

  60. MORE RESULTS Source 92

  61. MORE RESULTS Reference 93

  62. MORE RESULTS Warped source 94

  63. MORE RESULTS HDR 95

  64. MORE RESULTS Naïve fusion Our result 96

  65. MORE RESULTS Reference 97

  66. MORE RESULTS Source 98

  67. MORE RESULTS Reference 99

  68. MORE RESULTS Warped source 100

  69. MORE RESULTS HDR 101

  70. MORE RESULTS Naïve fusion Our result 102

  71. April 4-7, 2016 | Silicon Valley THAT’S ALL. O. Gallo A. Troccoli J. Hu K. Pulli J. Kautz

Recommend


More recommend