3D Reconstruction of Antique Stereo Pair Images

1
3D Reconstruction of Antique Stereo Pair Images

• Indriyati Atmosukarto, Anna Cavender, Chandrika
  Jayant

2
Objective

• What is the best way to reconstruct 3D scenes
  from antique stereo pair images?
• How do we handle over-segmentation in more
  complicated stereo pair images?

3
Related Work

• Single View Metrology Criminisi et al 99
• Uses only one image
• Manual input required
• Multiway Graph Cut Boykov, et al 98
• MinCut Algorithm Approximation with MRFs
• Slanted Surfaces Birchfield et al 99
• Uses Boykov and affine parameters
• Connected components

4
Our Work

• 1) Implement SVM
• 2) Implement Birchfield with Affine
• 3) Implement Graph Cut with User Constraints
• 4) Compare Results

5
Single View Metrology, Method,
Modeling

• Compute 3 vanishing points (x,y,z)
• Compute homography
• Define planes for 3D models
• Texture mapping to planes

6
Birchfield- Affine

• Graph cut code supplied
• Iteration method in paper
• Our idea
• Affine parameter estimation using feature
  matching (SIFT)

7
Graph Cut- User Constraints

• Use Graph Cut
• Run Connected Components
• Use user input to constrain regions and deal with
  oversegmentation
• Used lines and rectangles-
• Still very difficult even with user input!

8
Comparison

• Disparity ? 3d model
• 3d model ? disparity
• Slanted surfaces

9
Results
10
Tsukuba Stereo
11
SVM
12
Graph Cut-User Constraints
Graph Cut
Ours
13
3D Reconstruction of Disparity Map
14
Venus Stereo Graph Cut User Constraints
15
3D Reconstruction
16
Brussels Stereo- SVM
17
Graph Cut- User Constraints
Graph Cut Line Constraint
Rectangle Constraint
18
House Stereo
19
SVM
20
Graph Cut -User Constraints
Graph Cut Ground
Constraint More Constraints
21
Conclusion

• Antique photos hard to work with!
• Complex
• Noisy
• Dont know camera parameters
• Hard to compare methods
• Segmentation techniques must be improved

22
Future Plans

• Finish Affine Parameter Estimation
• Automated Rectangles for Oversegmentation Problem
• -First requires better user input method!
• Noise reduction for antique photos