mirror of
https://github.com/wassname/scikit-image.git
synced 2026-08-04 13:14:23 +08:00
Fixing indentation
This commit is contained in:
Ankit Agrawal
+54
-54
@@ -6,16 +6,16 @@ from scipy.spatial.distance import hamming
|
||||
|
||||
def _remove_border_keypoints(image, keypoints, dist):
|
||||
|
||||
width = image.shape[0]
|
||||
height = image.shape[1]
|
||||
width = image.shape[0]
|
||||
height = image.shape[1]
|
||||
|
||||
keypoints = keypoints[(dist < keypoints[:, 0]) & (keypoints[:, 0] < width - dist) &
|
||||
(dist < keypoints[:, 1]) & (keypoints[:, 1] < height - dist)]
|
||||
return keypoints
|
||||
keypoints = keypoints[(dist < keypoints[:, 0]) & (keypoints[:, 0] < width - dist) &
|
||||
(dist < keypoints[:, 1]) & (keypoints[:, 1] < height - dist)]
|
||||
return keypoints
|
||||
|
||||
|
||||
def brief(image, keypoints, descriptor_size=256, mode='normal', patch_size=49, sample_seed=1):
|
||||
"""Extract BRIEF Descriptor about given keypoints for a given image.
|
||||
"""Extract BRIEF Descriptor about given keypoints for a given image.
|
||||
|
||||
Parameters
|
||||
----------
|
||||
@@ -49,73 +49,73 @@ def brief(image, keypoints, descriptor_size=256, mode='normal', patch_size=49, s
|
||||
http://cvlabwww.epfl.ch/~lepetit/papers/calonder_eccv10.pdf
|
||||
|
||||
"""
|
||||
if np.squeeze(image).ndim == 3:
|
||||
image = rgb2gray(image)
|
||||
if np.squeeze(image).ndim == 3:
|
||||
image = rgb2gray(image)
|
||||
|
||||
keypoints = np.array(keypoints + 0.5, dtype=np.intp)
|
||||
keypoints = np.array(keypoints + 0.5, dtype=np.intp)
|
||||
|
||||
# Removing keypoints that are (patch_size / 2) distance from the image border
|
||||
keypoints = _remove_border_keypoints(image, keypoints, patch_size / 2)
|
||||
# Removing keypoints that are (patch_size / 2) distance from the image border
|
||||
keypoints = _remove_border_keypoints(image, keypoints, patch_size / 2)
|
||||
|
||||
descriptor = np.zeros((len(keypoints), descriptor_size), dtype=bool)
|
||||
descriptor = np.zeros((len(keypoints), descriptor_size), dtype=bool)
|
||||
|
||||
# Gaussian Low pass filtering with variance 2 to alleviate noise sensitivity
|
||||
image = gaussian_filter(image, 2)
|
||||
# Gaussian Low pass filtering with variance 2 to alleviate noise sensitivity
|
||||
image = gaussian_filter(image, 2)
|
||||
|
||||
# Sampling pairs of decision pixels in patch_size x patch_size window
|
||||
if mode == 'normal':
|
||||
np.random.seed(sample_seed)
|
||||
samples = np.round((patch_size / 5) * np.random.randn(descriptor_size * 8))
|
||||
samples = samples[(samples < (patch_size / 2)) & (samples > - (patch_size - 1) / 2)]
|
||||
first = (samples[: descriptor_size * 2]).reshape(descriptor_size, 2)
|
||||
second = (samples[descriptor_size * 2: descriptor_size * 4]).reshape(descriptor_size, 2)
|
||||
else:
|
||||
np.random.seed(sample_seed)
|
||||
samples = np.random.randint(-patch_size / 2, (patch_size / 2) + 1, (descriptor_size * 2, 2))
|
||||
first, second = np.split(samples, 2)
|
||||
# Sampling pairs of decision pixels in patch_size x patch_size window
|
||||
if mode == 'normal':
|
||||
np.random.seed(sample_seed)
|
||||
samples = np.round((patch_size / 5) * np.random.randn(descriptor_size * 8))
|
||||
samples = samples[(samples < (patch_size / 2)) & (samples > - (patch_size - 1) / 2)]
|
||||
first = (samples[: descriptor_size * 2]).reshape(descriptor_size, 2)
|
||||
second = (samples[descriptor_size * 2: descriptor_size * 4]).reshape(descriptor_size, 2)
|
||||
else:
|
||||
np.random.seed(sample_seed)
|
||||
samples = np.random.randint(-patch_size / 2, (patch_size / 2) + 1, (descriptor_size * 2, 2))
|
||||
first, second = np.split(samples, 2)
|
||||
|
||||
# Intensity comparison tests for building the descriptor
|
||||
for i in range(len(keypoints)):
|
||||
set_1 = first + keypoints[i]
|
||||
set_2 = second + keypoints[i]
|
||||
# Intensity comparison tests for building the descriptor
|
||||
for i in range(len(keypoints)):
|
||||
set_1 = first + keypoints[i]
|
||||
set_2 = second + keypoints[i]
|
||||
|
||||
for j in range(descriptor_size):
|
||||
if image[set_1[j, 0]][set_1[j, 1]] < image[set_2[j, 0]][set_2[j, 0]]:
|
||||
descriptor[i][j] = True
|
||||
for j in range(descriptor_size):
|
||||
if image[set_1[j, 0]][set_1[j, 1]] < image[set_2[j, 0]][set_2[j, 0]]:
|
||||
descriptor[i][j] = True
|
||||
|
||||
return descriptor
|
||||
return descriptor
|
||||
|
||||
|
||||
def hamming_distance(descriptor_1, descriptor_2):
|
||||
"""A dissimilarity measure used for matching keypoints in different images
|
||||
using binary feature descriptors like BRIEF etc.
|
||||
"""A dissimilarity measure used for matching keypoints in different images
|
||||
using binary feature descriptors like BRIEF etc.
|
||||
|
||||
Parameters
|
||||
----------
|
||||
descriptor_1 : ndarray with dtype bool
|
||||
Binary feature descriptor for keypoints in the first image.
|
||||
2D ndarray of dimensions (no_of_keypoints_in_image_1, descriptor_size)
|
||||
with value at an index (i, j) either being True or False representing
|
||||
the outcome of Intensity comparison about ith keypoint on jth decision
|
||||
pixel-pair.
|
||||
Binary feature descriptor for keypoints in the first image.
|
||||
2D ndarray of dimensions (no_of_keypoints_in_image_1, descriptor_size)
|
||||
with value at an index (i, j) either being True or False representing
|
||||
the outcome of Intensity comparison about ith keypoint on jth decision
|
||||
pixel-pair.
|
||||
descriptor_2 : ndarray with dtype bool
|
||||
Binary feature descriptor for keypoints in the second image.
|
||||
2D ndarray of dimensions (no_of_keypoints_in_image_2, descriptor_size)
|
||||
with value at an index (i, j) either being True or False representing
|
||||
the outcome of Intensity comparison about ith keypoint on jth decision
|
||||
pixel-pair.
|
||||
Binary feature descriptor for keypoints in the second image.
|
||||
2D ndarray of dimensions (no_of_keypoints_in_image_2, descriptor_size)
|
||||
with value at an index (i, j) either being True or False representing
|
||||
the outcome of Intensity comparison about ith keypoint on jth decision
|
||||
pixel-pair.
|
||||
|
||||
Returns
|
||||
-------
|
||||
distance : ndarray
|
||||
2D ndarray of dimensions (no_of_rows_in_descripto_1, no_of_rows_in_descripto_2)
|
||||
with value at an index (i, j) between the range [0, 1] representing the
|
||||
extent of dissimilarity between ith keypoint of in first image and jth
|
||||
keypoint in second image.
|
||||
2D ndarray of dimensions (no_of_rows_in_descripto_1, no_of_rows_in_descripto_2)
|
||||
with value at an index (i, j) between the range [0, 1] representing the
|
||||
extent of dissimilarity between ith keypoint of in first image and jth
|
||||
keypoint in second image.
|
||||
|
||||
"""
|
||||
distance = np.zeros((len(descriptor_1), len(descriptor_2)), dtype=float)
|
||||
for i in range(len(descriptor_1)):
|
||||
for j in range(len(descriptor_2)):
|
||||
distance[i, j] = hamming(descriptor_1[i][:], descriptor_2[j][:])
|
||||
return distance
|
||||
distance = np.zeros((len(descriptor_1), len(descriptor_2)), dtype=float)
|
||||
for i in range(len(descriptor_1)):
|
||||
for j in range(len(descriptor_2)):
|
||||
distance[i, j] = hamming(descriptor_1[i][:], descriptor_2[j][:])
|
||||
return distance
|
||||
|
||||
Reference in New Issue
Block a user