def __init__(self, clusters: ClusterResults, rates: List[float]) -> None:
self._results = [[] for _ in range(clusters.k())]
for clusterId, rate, path in zip(clusters.labels(), rates,
clusters.get_all_urls()):
self._results[clusterId].append({
'path': path,
'rating': rate,
'cluster': clusterId,
})
for cluster in self._results:
cluster.sort(key=lambda x: x['rating'], reverse=True)
def run(
self,
images: List[Url],
nfeatures: int = 0,
nOctaveLayers: int = 3,
contrastThreshold: float = 0.04,
edgeThreshold: float = 10,
sigma: float = 1.6,
ratio: float = 0.8,
similarity_metric: SimilarityMetric = SimilarityMetric.
INVERSE_DISTANCE,
damping: float = 0.5,
max_iter: int = 200,
convergence_iter: int = 15,
affinity: AffinityPropagationAffinity = AffinityPropagationAffinity.
EUCLIDEAN,
descriptor_matcher: DescriptorMatcher = DescriptorMatcher.FLANNBASED,
hue_bins: int = 180,
saturation_bins: int = 256,
value_bins: int = 256,
bandwidth: Optional[float] = None,
) -> ClusterResults:
results1 = SiftCluster2().run_cached(
images,
nfeatures=nfeatures,
nOctaveLayers=nOctaveLayers,
contrastThreshold=contrastThreshold,
edgeThreshold=edgeThreshold,
sigma=sigma,
ratio=ratio,
similarity_metric=similarity_metric,
damping=damping,
max_iter=max_iter,
convergence_iter=convergence_iter,
affinity=affinity,
descriptor_matcher=descriptor_matcher,
)
results2 = HistogramCluster().run_cached(
images,
hue_bins=hue_bins,
saturation_bins=saturation_bins,
value_bins=value_bins,
bandwidth=bandwidth,
)
labels1 = results1.labels()
labels2 = results2.labels()
k1 = results1.k()
cluster = [
self.combine(c1, c2, k1) for c1, c2 in zip(labels1, labels2)
]
self.remove_empty_clusters(cluster)
return ClusterResults(images, cluster)
def run(self, images: List[Url]) -> ClusterResults:
"""
Clusters images.
"""
results1 = self._sift.run(images)
results2 = self._hist.run(images)
labels1 = results1.labels()
labels2 = results2.labels()
k1 = results1.k()
cluster = [
self.combine(c1, c2, k1) for c1, c2 in zip(labels1, labels2)
]
self.remove_empty_clusters(cluster)
return ClusterResults(images, cluster)
def run(self, images: List[Url]) -> ClusterResults:
"""
Clusters images.
"""
results1 = self._hist.run(images)
cluster = [-1] * len(images)
for label1, urls1 in enumerate(results1.urls()):
results2 = self._sift.run(urls1)
for label2, urls2 in enumerate(results2.urls()):
label3 = self.combine(label1, label2, results1.k())
for url2 in urls2:
i = images.index(url2)
cluster[i] = label3
self.remove_empty_clusters(cluster)
return ClusterResults(images, cluster)
def run(self, images: List[Url]) -> ClusterResults:
"""
Creates descriptors of images.
Groups images together by how similar their descriptors are.
Returns a cluster ID for each set of descriptors.
"""
print("Creating descriptors from images....")
sds = SiftDescriptorSet(images)
print('Normalizing descriptors to unit vectors....')
sds.unit_normalize()
print('Similarity matrix....')
sm = SimilarityMatrix(sds.descriptors, self._similarity)
print('Scaling each row of the similarity matrix....')
sm.scale()
print('Clustering by affinity propagation....')
cluster = AffinityPropagation(random_state=0).fit_predict(
sm.matrix).tolist()
return ClusterResults(images, cluster)
def run(
self,
images: List[Url],
hue_bins: int = 180,
saturation_bins: int = 256,
value_bins: int = 256,
bandwidth: Optional[float] = None,
) -> ClusterResults:
"""
Clusters images.
"""
c = list()
for i, img in enumerate(images):
print("HISTOGRAM: %i / %i" % (i, len(images)))
hh = HsvHistogram(img)
histogram = HsvHistogram.scale(hh.hsv(hue_bins, saturation_bins, value_bins), hh.size())
c.append(histogram)
d = vstack(c)
print('CLUSTER: Mean Shift')
cluster = MeanShift(bandwidth=bandwidth).fit_predict(d).tolist()
return ClusterResults(images, cluster)
def run(
self,
images: List[Url],
nfeatures: int = 0,
nOctaveLayers: int = 3,
contrastThreshold: float = 0.04,
edgeThreshold: float = 10,
sigma: float = 1.6,
ratio: float = 0.8,
similarity_metric: SimilarityMetric = SimilarityMetric.
INVERSE_DISTANCE,
damping: float = 0.5,
max_iter: int = 200,
convergence_iter: int = 15,
affinity: AffinityPropagationAffinity = AffinityPropagationAffinity.
EUCLIDEAN,
descriptor_matcher: DescriptorMatcher = DescriptorMatcher.FLANNBASED,
) -> ClusterResults:
if not isinstance(similarity_metric, SimilarityMetric):
similarity_metric = SimilarityMetric(similarity_metric)
if not isinstance(affinity, AffinityPropagationAffinity):
affinity = AffinityPropagationAffinity(affinity)
if not isinstance(descriptor_matcher, DescriptorMatcher):
descriptor_matcher = DescriptorMatcher(descriptor_matcher)
list_of_images = list()
matrix = SimilarityMatrix.empty_matrix(len(images))
for url in images:
print("SIFT DESCRIPTORS: %s" % url)
keypoints, descriptors = cv2.xfeatures2d.SIFT_create(
nfeatures,
nOctaveLayers,
contrastThreshold,
edgeThreshold,
sigma,
).detectAndCompute(image=read_image(url), mask=None)
list_of_images.append(descriptors)
combo = list(
itertools.combinations_with_replacement(range(len(list_of_images)),
2))
if descriptor_matcher == DescriptorMatcher.FLANNBASED:
matcher = cv2.FlannBasedMatcher_create()
else:
matcher = cv2.BFMatcher_create()
for idx, (i, j) in enumerate(combo):
print("SIFT SIMILARITY: ( %i , %i ) %i / %i" %
(i, j, idx, len(combo)))
if i != j:
matches = matcher.knnMatch(queryDescriptors=list_of_images[i],
trainDescriptors=list_of_images[j],
k=2)
good = []
for m, n in matches:
if m.distance < ratio * n.distance:
good.append(m)
if similarity_metric == SimilarityMetric.INVERSE_DISTANCE:
inverse_distance = 0
for k in good:
inverse_distance += 1 - k.distance
if len(good) > 0:
matrix[i][j] = inverse_distance / len(good)
matrix[j][i] = inverse_distance / len(good)
elif similarity_metric == SimilarityMetric.COUNT:
matrix[i][j] = len(good)
matrix[j][i] = len(good)
else:
if similarity_metric == SimilarityMetric.INVERSE_DISTANCE:
matrix[i][i] = 1
elif similarity_metric == SimilarityMetric.COUNT:
matrix[i][i] = nfeatures
print('CLUSTER: AffinityPropagation')
cluster = AffinityPropagation(
damping=damping,
max_iter=max_iter,
convergence_iter=convergence_iter,
affinity=affinity.value,
random_state=0,
).fit_predict(matrix).tolist()
return ClusterResults(images, cluster)