1. 776 Computer Vision Jared Heinly Spring 2014 (slides borrowed from Jan-Michael Frahm, Svetlana Lazebnik, and others)

2. Image alignment Image from http://graphics.cs.cmu.edu/courses/15-463/2010_fall/

3. A look into the past http://blog.flickr.net/en/2010/01/27/a-look-into-the-past/

4. A look into the past Leningrad during the blockade http://komen-dant.livejournal.com/345684.html

5. Bing streetside images http://www.bing.com/community/blogs/maps/archive/2010/01/12/new-bing-maps- application-streetside-photos.aspx

6. Image alignment: Applications Panorama stitching Recognition of object instances

7. Image alignment: Challenges Small degree of overlap Occlusion, clutter Intensity changes

8. Image alignment Two families of approaches: o Direct (pixel-based) alignment Search for alignment where most pixels agree o Feature-based alignment Search for alignment where extracted features agree Can be verified using pixel-based alignment

9. 2D transformation models Similarity (translation, scale, rotation) Affine Projective (homography)

10. Let’s start with affine transformations Simple fitting procedure (linear least squares) Approximates viewpoint changes for roughly planar objects and roughly orthographic cameras Can be used to initialize fitting for more complex models

11. Fitting an affine transformation Assume we know the correspondences, how do we get the transformation?

12. Fitting an affine transformation Linear system with six unknowns Each match gives us two linearly independent equations: need at least three to solve for the transformation parameters

13. Fitting a plane projective transformation Homography: plane projective transformation (transformation taking a quad to another arbitrary quad)

14. Homography The transformation between two views of a planar surface The transformation between images from two cameras that share the same center

15. Application: Panorama stitching Source: Hartley & Zisserman

16. Fitting a homography Recall: homogeneous coordinates Converting to homogeneous image coordinates Converting from homogeneous image coordinates

17. Fitting a homography Recall: homogeneous coordinates Equation for homography: Converting to homogeneous image coordinates Converting from homogeneous image coordinates

18. Fitting a homography Equation for homography: 3 equations, only 2 linearly independent Unknown scale Scaled versions of same vector Rows of H Each element is a 3-vector Rows of H formed into 9x1 column vector

19. Direct linear transform H has 8 degrees of freedom (9 parameters, but scale is arbitrary) One match gives us two linearly independent equations Four matches needed for a minimal solution (null space of 8x9 matrix) More than four: homogeneous least squares

20. Robust feature-based alignment So far, we’ve assumed that we are given a set of “ground-truth” correspondences between the two images we want to align What if we don’t know the correspondences?

21. Robust feature-based alignment So far, we’ve assumed that we are given a set of “ground-truth” correspondences between the two images we want to align What if we don’t know the correspondences? ?

22. Robust feature-based alignment

23. Extract features

24. Robust feature-based alignment Extract features Compute putative matches

25. Alignment as fitting Least Squares Total Least Squares

26. Least Squares Line Fitting Data: (x 1, y 1 ), …, (x n, y n ) Line equation: y i = m x i + b Find ( m, b ) to minimize (x i, y i ) y=mx+b

27. Problem with “Vertical” Least Squares Not rotation-invariant Fails completely for vertical lines

28. Total Least Squares Distance between point (x i, y i ) and line ax+by=d (a 2 +b 2 =1): |ax i + by i – d| Find (a, b, d) to minimize the sum of squared perpendicular distances (x i, y i ) ax+by=d Unit normal: N=(a, b)

29. Least Squares: Robustness to Noise Least squares fit to the red points:

30. Least Squares: Robustness to Noise Least squares fit with an outlier: Problem: squared error heavily penalizes outliers

31. Robust Estimators General approach: find model parameters θ that minimize r i (x i, θ) – residual of i th point w.r.t. model parameters θ ρ – robust function with scale parameter σ The robust function ρ behaves like squared distance for small values of the residual u but saturates for larger values of u

32. Choosing the scale: Just right The effect of the outlier is minimized

33. The error value is almost the same for every point and the fit is very poor Choosing the scale: Too small

34. Choosing the scale: Too large Behaves much the same as least squares

35. RANSAC Robust fitting can deal with a few outliers – what if we have very many? Random sample consensus (RANSAC): Very general framework for model fitting in the presence of outliers Outline o Choose a small subset of points uniformly at random o Fit a model to that subset o Find all remaining points that are “close” to the model and reject the rest as outliers o Do this many times and choose the best model M. A. Fischler, R. C. Bolles. Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography. Comm. of the ACM, Vol 24, pp 381-395, 1981.Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography

36. RANSAC for line fitting example Source: R. Raguram

37. RANSAC for line fitting example Least-squares fit Source: R. Raguram

38. RANSAC for line fitting example 1.Randomly select minimal subset of points Source: R. Raguram

39. RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model Source: R. Raguram

40. RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function Source: R. Raguram

41. RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function 4.Select points consistent with model Source: R. Raguram

42. RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function 4.Select points consistent with model 5.Repeat hypothesize-and- verify loop Source: R. Raguram

43. 43 RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function 4.Select points consistent with model 5.Repeat hypothesize-and- verify loop Source: R. Raguram

44. 44 RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function 4.Select points consistent with model 5.Repeat hypothesize-and- verify loop Uncontaminated sample Source: R. Raguram

45. RANSAC for line fitting example 1.Randomly select minimal subset of points 2.Hypothesize a model 3.Compute error function 4.Select points consistent with model 5.Repeat hypothesize-and- verify loop Source: R. Raguram

46. RANSAC for line fitting Repeat N times: Draw s points uniformly at random Fit line to these s points Find inliers to this line among the remaining points (i.e., points whose distance from the line is less than t ) If there are d or more inliers, accept the line and refit using all inliers

47. Choosing the parameters Initial number of points s o Typically minimum number needed to fit the model Distance threshold t o Choose t so probability for inlier is p (e.g. 0.95) o Zero-mean Gaussian noise with std. dev. σ: t 2 =3.84σ 2 Number of samples N o Choose N so that, with probability p, at least one random sample is free from outliers (e.g. p=0.99) (outlier ratio: e) Source: M. Pollefeys

48. Choosing the parameters proportion of outliers e s5%10%20%25%30%40%50% 2235671117 33479111935 435913173472 54612172657146 64716243797293 748203354163588 8592644782721177 Source: M. Pollefeys Initial number of points s Typically minimum number needed to fit the model Distance threshold t Choose t so probability for inlier is p (e.g. 0.95) Zero-mean Gaussian noise with std. dev. σ: t 2 =3.84σ 2 Number of samples N Choose N so that, with probability p, at least one random sample is free from outliers (e.g. p=0.99) (outlier ratio: e)

49. Adaptively determining the number of samples Inlier ratio e is often unknown a priori, so pick worst case, e.g. 50%, and adapt if more inliers are found, e.g. 80% would yield e=0.2 Adaptive procedure: o N=∞, sample_count =0 o While N >sample_count Choose a sample and count the number of inliers If inlier ratio is highest of any found so far, set e = 1 – (number of inliers)/(total number of points) Recompute N from e: Increment the sample_count by 1 Source: M. Pollefeys

50. RANSAC pros and cons Pros o Simple and general o Applicable to many different problems o Often works well in practice Cons o Lots of parameters to tune o Doesn’t work well for low inlier ratios (too many iterations, or can fail completely) o Can’t always get a good initialization of the model based on the minimum number of samples

51. Alignment as fitting Previous lectures: fitting a model to features in one image Find model M that minimizes M xixi

52. Alignment as fitting Previous lectures: fitting a model to features in one image Alignment: fitting a model to a transformation between pairs of features (matches) in two images Find model M that minimizes Find transformation T that minimizes M xixi T xixi xixi '

53. Robust feature-based alignment Extract features Compute putative matches Loop: o Hypothesize transformation T

54. Robust feature-based alignment Extract features Compute putative matches Loop: o Hypothesize transformation T o Verify transformation (search for other matches consistent with T)

55. Robust feature-based alignment Extract features Compute putative matches Loop: o Hypothesize transformation T o Verify transformation (search for other matches consistent with T)

56. RANSAC The set of putative matches contains a very high percentage of outliers RANSAC loop: 1.Randomly select a seed group of matches 2.Compute transformation from seed group 3.Find inliers to this transformation 4.If the number of inliers is sufficiently large, re- compute least-squares estimate of transformation on all of the inliers Keep the transformation with the largest number of inliers

57. RANSAC example: Translation Putative matches

58. RANSAC example: Translation Select one match, count inliers

59. RANSAC example: Translation Select one match, count inliers

60. RANSAC example: Translation Select translation with the most inliers

61. Summary Detect feature points (to be covered later) Describe feature points (to be covered later) Match feature points (to be covered later) RANSAC o Transformation model (affine, homography, etc.)

62. Advice Use the minimum model necessary to define the transformation on your data o Affine in favor of homography o Homography in favor of 3D transform (not discussed yet) o Make selection based on your data Fewer RANSAC iterations Fewer degrees of freedom  less ambiguity or sensitivity to noise

63. Generating putative correspondences ?

64. Need to compare feature descriptors of local patches surrounding interest points ( ) = ? feature descriptor ?

65. Simplest descriptor: vector of raw intensity values How to compare two such vectors? o Sum of squared differences (SSD) Not invariant to intensity change o Normalized correlation Invariant to affine intensity change Feature descriptors

66. Disadvantage of intensity vectors as descriptors Small deformations can affect the matching score a lot

67. Descriptor computation: o Divide patch into 4x4 sub-patches o Compute histogram of gradient orientations (8 reference angles) inside each sub-patch o Resulting descriptor: 4x4x8 = 128 dimensions Feature descriptors: SIFT David G. Lowe. "Distinctive image features from scale-invariant keypoints.” IJCV 60 (2), pp. 91-110, 2004."Distinctive image features from scale-invariant keypoints.”

68. Descriptor computation: o Divide patch into 4x4 sub-patches o Compute histogram of gradient orientations (8 reference angles) inside each sub-patch o Resulting descriptor: 4x4x8 = 128 dimensions Advantage over raw vectors of pixel values o Gradients less sensitive to illumination change o Pooling of gradients over the sub-patches achieves robustness to small shifts, but still preserves some spatial information Feature descriptors: SIFT David G. Lowe. "Distinctive image features from scale-invariant keypoints.” IJCV 60 (2), pp. 91-110, 2004."Distinctive image features from scale-invariant keypoints.”

69. Feature matching ? Generating putative matches: for each patch in one image, find a short list of patches in the other image that could match it based solely on appearance

70. Feature space outlier rejection How can we tell which putative matches are more reliable? Heuristic: compare distance of nearest neighbor to that of second nearest neighbor o Ratio of closest distance to second-closest distance will be high for features that are not distinctive David G. Lowe. "Distinctive image features from scale-invariant keypoints.” IJCV 60 (2), pp. 91-110, 2004."Distinctive image features from scale-invariant keypoints.” Threshold of 0.8 provides good separation