ai_ct_scans.keypoint_alignment module

Module to provide functions to perform keypoint detection alignment.

ai_ct_scans.keypoint_alignment.align_image(image, reference, detector='SIFT')

Align an image to a reference image using keypoint detection.

Parameters

  • image (np.ndarray) – A 2D greyscale image to be aligned.

  • reference (np.ndarray) – A 2D greyscale image that to be used as the reference.

  • detector (str) – The keypoint detector to use. Should be one of [‘SIFT’, ‘ORB’].

Returns

A 2D greyscale image that is the result of the alignment.

Return type

(np.ndarray)

ai_ct_scans.keypoint_alignment.find_homography(key_points_1, key_points_2, good_matches)

Estimate the transformation that maps one set of keypoints to another.

Parameters

  • key_points_1 (list of cv2.KeyPoint) – The keypoints extracted from the source image.

  • key_points_2 (list of cv2.KeyPoint) – The keypoints extracted from the destination image.

  • good_matches (list of cv2.DMatch) – Matched descriptors.

Returns

A transformation matrix that maps the first set of keypoints to the second. (np.ndarray): Mask of inlier and outlier points used to estimate transformation.

Return type

(np.ndarray)

ai_ct_scans.keypoint_alignment.get_keypoints_and_descriptors(image, detector='SIFT')

Get keypoints and compute corresponding descriptors from an image.

Parameters

  • image (np.ndarray) – A 2D greyscale image.

  • detector (str) – The keypoint detector to use. Should be one of [‘SIFT’, ‘ORB’].

Returns

List of keypoints as cv2.KeyPoint (list): Descriptor corresponding to the returned keypoints

Return type

(list)

ai_ct_scans.keypoint_alignment.match_descriptors(descriptors_1, descriptors_2)

Match descriptors that have come from two separate images.

Parameters

  • descriptors_1 (np.ndarray) – Descriptors that correspond to keypoints extracted from the first image.

  • descriptors_2 (np.ndarray) – Descriptors that correspond to keypoints extracted from the first image.

Returns

Matched descriptors.

Return type

(list of cv2.DMatch)

ai_ct_scans.keypoint_alignment.sieve_matches_lowe(matches)

Apply Lowe’s ratio test to retain only good matches.

Parameters

matches – (list of cv2.DMatch): Matched descriptors.

Returns

Matched descriptors that passed the test.

Return type

(list of cv2.DMatch)