Computes the descriptors for a set of keypoints detected in an image or image set.

[descriptors, keypoints] = extractor.compute(img, keypoints)
[descriptors, keypoints] = extractor.compute(imgs, keypoints)

Input

• img Image (first variant).
• imgs Image set (second variant), cell array of images.
• keypoints Input collection of keypoints. Keypoints for which a descriptor cannot be computed are removed. Sometimes new keypoints can be added, for example: cv.SIFT duplicates keypoint with several dominant orientations (for each orientation). In the first variant, this is a struct-array of detected keypoints. In the second variant, it is a cell-array, where keypoints{i} is a set of keypoints detected in images{i} (a struct-array like before). Each keypoint is a struct with the following fields:
  • pt coordinates of the keypoint [x,y]
  • size diameter of the meaningful keypoint neighborhood
  • angle computed orientation of the keypoint (-1 if not applicable); it's in [0,360) degrees and measured relative to image coordinate system (y-axis is directed downward), i.e in clockwise.
  • response the response by which the most strong keypoints have been selected. Can be used for further sorting or subsampling.
  • octave octave (pyramid layer) from which the keypoint has been extracted.
  • class_id object class (if the keypoints need to be clustered by an object they belong to).

Output

• descriptors Computed descriptors. In the second variant of the method descriptors{i} are descriptors computed for a keypoints{i}. Row j in descriptors (or descriptors{i}) is the descriptor for j-th keypoint.
• keypoints Optional output with possibly updated keypoints.