def main():
parser = argparse.ArgumentParser()
parser.add_argument(
'in_dissim_matrix',
default='../dissim_matrix_hybrid.pickle',
help='The dissimilarity-Matrix to be used as a pickled Numpy-Array')
parser.add_argument(
'in_cores',
default='../h_cores_as_indexes.pickle',
help='file path of cores file: "../h_cores_as_indexes.pickle"')
parser.add_argument('in_values')
parser.add_argument('random_init')
args = parser.parse_args()
f = open(args.in_dissim_matrix)
distances = pickle.load(f)
f.close()
cores = loadtxt(args.in_cores)
if args.random_init != "random":
init = getInit(distances, cores)
set_printoptions(threshold='nan')
set_printoptions(linewidth=900000000)
result = kmedoids(distance=distances, npass=20, nclusters=len(cores))
#result = kmedoids(distance = distances, initialid = init)
print args.in_values + " " + str(len(cores)) + " " + str(
result[1]) + " " + str(result[0])
def kMedoids(args): """ Do k-medoids clustering on a distance matrix. @param args: A tuple of the form (dist_matrix, k, n_passes) @return: The result tuple returned by Pycluster.kmedoids """ dist_matrix, k, n_passes = args return kmedoids(dist_matrix, k, n_passes)
def main():
parser = argparse.ArgumentParser()
parser.add_argument('in_dissim_matrix', default='../dissim_matrix_hybrid.pickle', help='The dissimilarity-Matrix to be used as a pickled Numpy-Array')
parser.add_argument('in_cores', default='../h_cores_as_indexes.pickle', help='file path of cores file: "../h_cores_as_indexes.pickle"')
parser.add_argument('in_values')
parser.add_argument('random_init')
args = parser.parse_args()
f = open(args.in_dissim_matrix)
distances = pickle.load(f)
f.close()
cores = loadtxt(args.in_cores)
if args.random_init != "random":
init = getInit( distances, cores )
set_printoptions(threshold='nan')
set_printoptions(linewidth=900000000)
result = kmedoids(distance = distances, npass=20, nclusters=len(cores))
#result = kmedoids(distance = distances, initialid = init)
print args.in_values + " "+str(len(cores))+" "+str(result[1])+" "+str(result[0])
inorm1 = inorm[DLi:DHi + 1]
inorm1 = np.swapaxes(inorm1, 0, 1)
inorm1 = inorm1[ELi:EHi + 1]
inorm1 = np.swapaxes(inorm1, 0, 1)
difmat = inorm1[1] - i1norm1[1]
difmat2 = difmat**2
RMSD = (np.average(difmat2))**0.5
pRMSD = RMSD * 100
rmsdmat1.append(pRMSD)
rmsdmat.append(rmsdmat1)
rmsdmat = np.array(rmsdmat)
rmsdmat = 1 - rmsdmat
np.savetxt("DifferenceMatrix.csv", rmsdmat, delimiter=",")
clusterid, error, nfound = kmedoids(rmsdmat, nclusters=c, npass=10)
print "Error = ", error
print "Found this configuration ", nfound, " out of 10 times"
with open("Clusters.csv", "wb") as csvfile:
writer = csv.writer(csvfile, delimiter=",")
writer.writerow(["item #", "File Name", "Cluster #"])
writer.writerow([""])
for i in range(0, len(clusterid)):
print i, files[i], " cluster = ", clusterid[i]
writer.writerow([i, files[i], clusterid[i]])
writer.writerow([""])
writer.writerow([
"Note : The cluster number is defined as the item number of the centroid of the cluster."
])
print "\n" + "The cluster number is defined as the item number of the centroid of the cluster."
def get_surface_sources(surface, space=5, distance='euclidean', remains=None):
"""get sources in volume
Parameters
----------
surface : Surface object
space : float
The distance between sources
distance : 'euclidean' | 'dijkstra' | 'continuous'
The distance used to compute distance on surface
remains : None | int
The number of sources that we want to keep
Returns
-------
src : SourceSpaces object
-------
Author : Alexandre Fabre
"""
if remains is None:
remains, removes = get_number_sources(surface, space=space,
surface=True)
else:
# avoid to have a incorrect number of sources
remains = max(0, min(surface.pos_length, remains))
removes = surface.pos_length - remains
if remains == 0:
raise ValueError('Error, 0 source created')
if removes == 0:
# all points are sources
# logger.info('all points are remained')
centroids_id = np.arange(remains)
inuse = np.ones(surface.pos_length, dtype=int)
else:
# connectivity of neighbors points
n_neighbors = min(50, surface.pos_length)
# get the matrix that identify neighbors points
knn_graph = kneighbors_graph(surface.pos, n_neighbors, include_self=False)
# ward criterion is adapted for a surface clustering
model = AgglomerativeClustering(linkage='ward', connectivity=knn_graph,
n_clusters=remains)
# compute clusters
model.fit(surface.pos)
# get cluster labels
cluster_id = model.labels_
# get the distance between points on the surface with Dijkstra or continuous
# if distance is euclidean, it just computes euclidean distances between points
distance = surf_m.get_surf_distance(surface.pos, surface.triangles,
distance=distance)
# clusters give by AgglomerativeClustering are initial clusters for k-medoids
# for k-medoids, the centroid is a point in a cluster
# k-medoids method return clusters that are identified by the index of their centroid point
cluster_id, _, _ = kmedoids(distance, nclusters=remains, npass=1,
initialid=cluster_id)
# get the index of centroids
centroids_id = np.unique(cluster_id)
inuse = np.zeros(surface.pos_length)
inuse[centroids_id] = 1
inuse = inuse.astype(int) # Need to be int
# must be converted to meters and transorm to numpy array
rr = surface.pos * 1e-3
# Change index for hemi
if surface.hemi=='lh':
Id = 101
elif surface.hemi=='rh':
Id = 102
src = [{'rr': rr, 'coord_frame': np.array((FIFF.FIFFV_COORD_MRI,), np.int32), 'type': 'surf', 'id': Id,
'np': surface.pos_length, 'nn': surface.normals, 'inuse': inuse, 'nuse': remains, 'dist': None,
'ntri': surface.triangles_length, 'nearest': None, 'use_tris': None, 'nuse_tris': 0,
'vertno': centroids_id, 'patch_inds': None, 'tris': surface.triangles, 'dist_limit': None, 'pinfo': None,
'nearest_dist': None, 'removes': removes}]
src = SourceSpaces(src)
return src
def main():
p = opt.ArgumentParser(description="""
Constructs a dictionary for image representation based on histograms of codeblocks
(Gabor wavelet local descriptors) over larger neighborhoods.
The dictionary is built from a set of images given as a list in an input file.
""")
p.add_argument('img_path', action='store', help='path to image files - all images in the folder will be used')
p.add_argument('img_ext', action='store', help='extension of the image files (e.g. "jpg" or "png") - NO DOT!')
p.add_argument('l0_model', action='store', help='level-0 codebook model file')
p.add_argument('out_file', action='store', help='resulting model file name')
p.add_argument('codebook_size', action='store', help='codebook size', type=int)
p.add_argument('-w', '--window', action='store', help='local window size (default: 512)', type=int, default=512)
args = p.parse_args()
#---------
# data
data_path = args.img_path
img_ext = args.img_ext
wnd_size = args.window
with ModelPersistence(args.l0_model, 'r', format='pickle') as mp:
l0_model = mp
img_files = glob.glob(data_path + '/*.' + img_ext)
if len(img_files) == 0:
return
#---------
# Gabor
tmp = np.array([0.0, np.pi / 4.0, np.pi / 2.0, 3.0 * np.pi / 4.0], dtype=np.double)
tmp2 = np.array([3.0 / 4.0, 3.0 / 8.0, 3.0 / 16.0], dtype=np.double)
tmp3 = np.array([1.0, 2 * np.sqrt(2.0)], dtype=np.double)
local_descriptor = GaborDescriptor(theta=tmp, freq=tmp2, sigma=tmp3)
## Process:
sys.stdout = os.fdopen(sys.stdout.fileno(), 'w', 0) # unbuferred output
desc_vectors = [] # a list of local descriptor vectors
print('Computing level 0 coding...')
desc_vectors = \
Parallel(n_jobs=cpu_count()) \
( delayed(worker)(img_name, local_descriptor, wnd_size, l0_model) for img_name in img_files )
print('OK')
print('Vector quantization:')
print('-prepare...')
X = np.vstack(desc_vectors) # each row is a histogram
np.save('X_bag_level1.dat', X)
print('-compute pairwise distances...')
n = X.shape[0]
pdist = Parallel(n_jobs=cpu_count()) ( delayed(worker_chisq_M)(X[i,:], X[i+1:n,:]) for i in np.arange(0, n-1) )
# make the list flat:
pdist = np.array(list(itertools.chain.from_iterable(pdist)))
#for i in np.arange(0, X.shape[0]-1):
# for j in np.arange(i+1, X.shape[0]):
# pdist.append(dist.chisq(X[i,:], X[j,:]))
pdist = np.array(pdist)
np.save('X_pdist_level1.data.npy', pdist)
print('-cluster (k-medoids)...')
meds = kmedoids(pdist, nclusters=args.codebook_size, npass=20)
labels = np.unique(meds[0]) # also the indexes of vectors from X that became cluster centers (medoids)
vq = {}
vq['cluster_centers_'] = X[labels, :]
vq['labels_'] = labels
vq['distance'] = 'chisq'
print('OK')
print('Saving model...', end='')
# compute the average distance and std.dev. of the points in each cluster:
avg_dist = np.zeros(args.codebook_size)
sd_dist = np.zeros(args.codebook_size)
for k in range(0, args.codebook_size):
idx = np.where(meds[0] == labels[k])[0]
d = []
for i in idx:
d.append(dist.chisq(X[i,:], vq['cluster_centers_'][k,:]))
avg_dist[k] = np.array(d).mean()
sd_dist[k] = np.array(d).std()
print('K-medoids summary:');
print('-avg. dist: ', avg_dist)
print('-std. dev. dist: ', sd_dist)
with ModelPersistence(args.out_file, 'c', format='pickle') as d:
d['codebook'] = vq
d['avg_dist_to_centroid'] = avg_dist
d['stddev_dist_to_centroid'] = sd_dist
print('OK')
return
def main():
p = opt.ArgumentParser(description="""
Constructs a dictionary for image representation based on histograms of codeblocks
(Gabor wavelet local descriptors) over larger neighborhoods.
The dictionary is built from a set of images given as a list in an input file.
""")
p.add_argument(
'img_path',
action='store',
help='path to image files - all images in the folder will be used')
p.add_argument(
'img_ext',
action='store',
help='extension of the image files (e.g. "jpg" or "png") - NO DOT!')
p.add_argument('l0_model',
action='store',
help='level-0 codebook model file')
p.add_argument('out_file',
action='store',
help='resulting model file name')
p.add_argument('codebook_size',
action='store',
help='codebook size',
type=int)
p.add_argument('-w',
'--window',
action='store',
help='local window size (default: 512)',
type=int,
default=512)
args = p.parse_args()
#---------
# data
data_path = args.img_path
img_ext = args.img_ext
wnd_size = args.window
with ModelPersistence(args.l0_model, 'r', format='pickle') as mp:
l0_model = mp
img_files = glob.glob(data_path + '/*.' + img_ext)
if len(img_files) == 0:
return
#---------
# Gabor
tmp = np.array([0.0, np.pi / 4.0, np.pi / 2.0, 3.0 * np.pi / 4.0],
dtype=np.double)
tmp2 = np.array([3.0 / 4.0, 3.0 / 8.0, 3.0 / 16.0], dtype=np.double)
tmp3 = np.array([1.0, 2 * np.sqrt(2.0)], dtype=np.double)
local_descriptor = GaborDescriptor(theta=tmp, freq=tmp2, sigma=tmp3)
## Process:
sys.stdout = os.fdopen(sys.stdout.fileno(), 'w', 0) # unbuferred output
desc_vectors = [] # a list of local descriptor vectors
print('Computing level 0 coding...')
desc_vectors = \
Parallel(n_jobs=cpu_count()) \
( delayed(worker)(img_name, local_descriptor, wnd_size, l0_model) for img_name in img_files )
print('OK')
print('Vector quantization:')
print('-prepare...')
X = np.vstack(desc_vectors) # each row is a histogram
np.save('X_bag_level1.dat', X)
print('-compute pairwise distances...')
n = X.shape[0]
pdist = Parallel(n_jobs=cpu_count())(
delayed(worker_chisq_M)(X[i, :], X[i + 1:n, :])
for i in np.arange(0, n - 1))
# make the list flat:
pdist = np.array(list(itertools.chain.from_iterable(pdist)))
#for i in np.arange(0, X.shape[0]-1):
# for j in np.arange(i+1, X.shape[0]):
# pdist.append(dist.chisq(X[i,:], X[j,:]))
pdist = np.array(pdist)
np.save('X_pdist_level1.data.npy', pdist)
print('-cluster (k-medoids)...')
meds = kmedoids(pdist, nclusters=args.codebook_size, npass=20)
labels = np.unique(
meds[0]
) # also the indexes of vectors from X that became cluster centers (medoids)
vq = {}
vq['cluster_centers_'] = X[labels, :]
vq['labels_'] = labels
vq['distance'] = 'chisq'
print('OK')
print('Saving model...', end='')
# compute the average distance and std.dev. of the points in each cluster:
avg_dist = np.zeros(args.codebook_size)
sd_dist = np.zeros(args.codebook_size)
for k in range(0, args.codebook_size):
idx = np.where(meds[0] == labels[k])[0]
d = []
for i in idx:
d.append(dist.chisq(X[i, :], vq['cluster_centers_'][k, :]))
avg_dist[k] = np.array(d).mean()
sd_dist[k] = np.array(d).std()
print('K-medoids summary:')
print('-avg. dist: ', avg_dist)
print('-std. dev. dist: ', sd_dist)
with ModelPersistence(args.out_file, 'c', format='pickle') as d:
d['codebook'] = vq
d['avg_dist_to_centroid'] = avg_dist
d['stddev_dist_to_centroid'] = sd_dist
print('OK')
return
data = squareform(data)
else:
print "\n>>> Loading provided distance matrix..."
data = np.loadtxt('%s' % input_dist)
if write_dist:
print "\n>>> Writing distance_matrix.txt..."
np.savetxt('distance_matrix.txt', data, fmt='%10.3f')
if clus_min != clus_max:
#Compute silhouette score for k=kmin -> k=kmax:
print "\n>>> Determining optimal number of clusters using silhouette score..."
sil = numpy.zeros(clus_max - clus_min + 1)
for i in range(clus_min, clus_max + 1):
idx, error, nfound = kmedoids(data, nclusters=i, npass=passes)
sil[i - clus_min] = silhouette_score(data,
idx,
metric='precomputed',
sample_size=None,
random_state=None)
clus_num = numpy.argmax(sil) + clus_min
print "\tOptimal number of clusters: %s" % clus_num
else:
sil = sil = numpy.zeros(clus_max - clus_min + 1)
clus_num = clus_min
print "\n>>> Requested %s clusters" % clus_num
#K-Medoids clustering
idx, error, nfound = kmedoids(data, nclusters=clus_num, npass=passes)
