1 / 52

More on Features

More on Features. Digital Visual Effects, Spring 2006 Yung-Yu Chuang 2006/3/22. with slides by Trevor Darrell Cordelia Schmid , David Lowe, Darya Frolova, Denis Simakov , Robert Collins and Jiwon Kim. Announcements.

Télécharger la présentation

More on Features

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. More on Features Digital Visual Effects, Spring 2006 Yung-Yu Chuang 2006/3/22 with slides by Trevor DarrellCordelia Schmid, David Lowe, Darya Frolova, Denis Simakov, Robert Collins and Jiwon Kim

  2. Announcements • Project #1 was due at 11:59pm this Saturday. Please send to TAs a mail including a link to a zip file of your submission. • Project #2 handout will be available on the web on Sunday.

  3. Outline • Harris corner detector • SIFT • SIFT extensions • MSOP

  4. Three components for features • Feature detection • Feature description • Feature matching

  5. Harris corner detector

  6. Harris corner detector • Consider all small shifts by Taylor’s expansion

  7. Harris corner detector Equivalently, for small shifts [u,v] we have a bilinear approximation: , where M is a 22 matrix computed from image derivatives:

  8. Visualize quadratic functions

  9. Visualize quadratic functions

  10. Visualize quadratic functions

  11. Visualize quadratic functions

  12. Harris corner detector (matrix form)

  13. Harris corner detector Intensity change in shifting window: eigenvalue analysis 1, 2 – eigenvalues of M direction of the fastest change Ellipse E(u,v) = const direction of the slowest change (max)-1/2 (min)-1/2

  14. Harris corner detector Classification of image points using eigenvalues of M: 2 edge 2 >> 1 Corner 1 and 2 are large,1 ~ 2;E increases in all directions 1 and 2 are small;E is almost constant in all directions edge 1 >> 2 flat 1

  15. Harris corner detector Measure of corner response: (k – empirical constant, k = 0.04-0.06)

  16. Summary of Harris detector

  17. Now we know where features are • But, how to match them? • What is the descriptor for a feature? The simplest solution is the intensities of its spatial neighbors. This might not be robust to brightness change or small shift/rotation.

  18. Harris Detector: Some Properties • Rotation invariance Ellipse rotates but its shape (i.e. eigenvalues) remains the same Corner response R is invariant to image rotation

  19. Harris Detector: Some Properties • But: non-invariant to image scale! All points will be classified as edges Corner !

  20. Scale invariant detection • The problem: how do we choose corresponding circles independently in each image? • Aperture problem

  21. SIFT (Scale Invariant Feature Transform)

  22. descriptor detector ( ) local descriptor SIFT stages: • Scale-space extrema detection • Keypoint localization • Orientation assignment • Keypoint descriptor matching A 500x500 image gives about 2000 features

  23. 1. Detection of scale-space extrema • For scale invariance, search for stable features across all possible scales using a continuous function of scale, scale space. • SIFT uses DoG filter for scale space because it is efficient and as stable as scale-normalized Laplacian of Gaussian.

  24. Scale space  doubles for the next octave K=2(1/s)

  25. Detection of scale-space extrema

  26. Keypoint localization X is selected if it is larger or smaller than all 26 neighbors

  27. 2. Accurate keypoint localization • Reject points with low contrast and poorly localized along an edge • Fit a 3D quadratic function for sub-pixel maxima

  28. Accurate keypoint localization • Change sample point if offset is larger than 0.5 • Throw out low contrast (<0.03)

  29. Eliminating edge responses Let Keep the points with r=10

  30. Maxima in D

  31. Remove low contrast and edges

  32. 3. Orientation assignment • By assigning a consistent orientation, the keypoint descriptor can be orientation invariant. • For a keypoint, L is the image with the closest scale, orientation histogram

  33. Orientation assignment

  34. Orientation assignment

  35. Orientation assignment

  36. Orientation assignment

  37. SIFT descriptor

  38. 4. Local image descriptor • Thresholded image gradients are sampled over 16x16 array of locations in scale space • Create array of orientation histograms (w.r.t. key orientation) • 8 orientations x 4x4 histogram array = 128 dimensions • Normalized, clip values larger than 0.2, renormalize σ=0.5*width

  39. SIFT extensions

  40. PCA

  41. PCA-SIFT • Only change step 4 • Pre-compute an eigen-space for local gradient patches of size 41x41 • 2x39x39=3042 elements • Only keep 20 components • A more compact descriptor

  42. GLOH (Gradient location-orientation histogram) SIFT 17 location bins 16 orientation bins Analyze the 17x16=272-d eigen-space, keep 128 components

  43. Multi-Scale Oriented Patches • Simpler than SIFT. Designed for image matching. [Brown, Szeliski, Winder, CVPR’2005] • Feature detector • Multi-scale Harris corners • Orientation from blurred gradient • Geometrically invariant to rotation • Feature descriptor • Bias/gain normalized sampling of local patch (8x8) • Photometrically invariant to affine changes in intensity

  44. Multi-Scale Harris corner detector • Image stitching is mostly concerned with matching images that have the same scale, so sub-octave pyramid might not be necessary.

  45. Multi-Scale Harris corner detector gradient of smoother version Corner detection function: Pick local maxima of 3x3 and larger than 10

  46. Orientation assignment • Orientation = blurred gradient

  47. MOPS descriptor vector • Scale-space position (x, y, s) + orientation () • 8x8 oriented patch sampled at 5 x scale. See the Technical Report for more details. • Bias/gain normalisation: I’ = (I –)/ 40 pixels 8 pixels

  48. Detections at multiple scales

  49. Feature matching • Exhaustive search • for each feature in one image, look at all the other features in the other image(s) • Hashing • compute a short descriptor from each feature vector, or hash longer descriptors (randomly) • Nearest neighbor techniques • k-trees and their variants (Best Bin First)

  50. Wavelet-based hashing • Compute a short (3-vector) descriptor from an 8x8 patch using a Haar “wavelet” • Quantize each value into 10 (overlapping) bins (103 total entries) • [Brown, Szeliski, Winder, CVPR’2005]

More Related