network_name=args.network_name,
checkpoint_path=args.checkpoint,
batch_size=args.batch_size,
num_classes=args.num_classes,
preproc_func_name=args.preproc_func,
preproc_threads=args.num_preproc_threads)
# Print the network summary, use these layer names for feature extraction
#feature_extractor.print_network_summary()
# Feature extraction example using a filename queue to feed images
feature_dataset = feature_extraction_queue(feature_extractor,
args.image_path, layer_names,
args.batch_size,
args.num_classes)
# print(type(feature_dataset))
# for i in feature_dataset:
# print(i)
data = feature_dataset['resnet_v2_101/logits']
data = np.reshape(data, [data.shape[0], data.shape[3]])
np.save(args.out_file, data)
# Write features to disk as HDF5 file
utils.write_hdf5('features.h5', layer_names, feature_dataset)
print("Successfully written features to: {}".format(args.out_file))
# Close the threads and close session.
feature_extractor.close()
print("Finished.")
label3 = label2.split('_')
# Get the features for this label.
temp = features[index_label]
temp0 = np.append(temp, int(label3[0]) - 1) # category
temp1 = np.append(temp0, int(label3[1]) - 1) # object class
temp2 = np.append(temp1, label3[2]) # session
temp3 = np.append(temp2, label3[3][:-4]) # image id within the object class
features_array[index_label] = temp3
feature_dataset[layer_names[0]] = features_array
# End: Modifications (Vadym Gryshchuk).
# Write features to disk as HDF5 file
utils.write_hdf5(args.out_file, layer_names, feature_dataset)
print("Successfully written features to: {}".format(args.out_file))
# Close the threads and close session.
feature_extractor.close()
print("Finished.")
# Start: Modifications (Vadym Gryshchuk).
f = h5py.File(args.out_file, 'r')
dataset = f[args.layer_names][...]
np.savetxt(args.out_features_file, dataset, delimiter=',')
np.save
print("Features are written also to: {}".format(args.out_features_file))
print("Done.")
