def bow_train(featfunc, opts, canon_opts, params):
descs = []
files = dataset.training_files(opts['num_train_images'])
for img_file, depth_file in print_progress(files):
img, segmask = canonize(img_file, depth_file, canon_opts, params)
if featfunc == 'daisy':
h = daisy(img_as_float(img), **opts['feature_opts'])
if featfunc == 'jet':
h = jet(img_as_float(img), **opts['feature_opts'])
segmask = sp.misc.imresize(segmask, h.shape[:2])
for i in range(h.shape[0]):
for j in range(h.shape[1]):
if segmask[i,j] != 0:
descs.append(h[i,j,:].flatten())
descs = [descs[i] for i in range(0, len(descs), opts['feature_step'])]
descs = np.vstack(tuple(descs))
print '# K-means clustering of %i features of dimensionality %i'%(descs.shape[0], descs.shape[1])
kmeans = KMeans(opts['num_clusters'], n_jobs=options.num_threads)
kmeans.fit(descs)
return kmeans
def feature_extraction(img_file, depth_file, opts, params):
img, segmask = canonize(img_file, depth_file, opts['canonization'], params)
features = np.array([], dtype=float)
if 'region_properties' in opts:
opts_ = opts['region_properties']
if 'grid' in opts_:
grid = opts_['grid']
segmask[segmask!=0] = 1
cellNum = 0
cellHeight = segmask.shape[0]/grid[0]
cellWidth = segmask.shape[1]/grid[1]
for i in range(grid[0]):
for j in range(grid[1]):
cellNum += 1
t = cellHeight*i
b = cellHeight*(i+1)
l = cellWidth*j
r = cellWidth*(j+1)
segmask[t:b, l:r] = segmask[t:b, l:r] * cellNum
props = regionprops(segmask.astype(int), properties=opts_['properties'], intensity_image=img)
featureList = []
for i in range(len(props)):
for p in opts_['properties']:
featureList.append(np.array(props[i][p]).flatten())
features = np.append(features, np.hstack(tuple(featureList)))
if 'raw_pixels' in opts:
img_ = sp.misc.imresize(img, opts['raw_pixels']['img_scale'])
features = np.append(features, img_.flatten())
if 'daisy_bow' in opts:
f = bow(img, segmask, daisy, opts['daisy_bow'], params['daisy_bow_kmeans'])
features = np.append(features, f)
if 'jet_bow' in opts:
f = bow(img, segmask, jet, opts['jet_bow'], params['jet_bow_kmeans'])
features = np.append(features, f)
return features
def output_img(filename, outdir, output_filename, params): img_file, depth_file = filename img, _ = canonize(img_file, depth_file, options.canonization, params) sp.misc.imsave(os.path.join(outdir, output_filename), img)
