keras.backend.round(keras.backend.clip(y_pred, 0, 1)))
precision = true_positives / (predicted_positives +
keras.backend.epsilon())
return precision
precision = precision(y_true, y_pred)
recall = recall(y_true, y_pred)
return 2 * ((precision * recall) /
(precision + recall + keras.backend.epsilon()))
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('-c',
'--config',
help="path to configuration file",
default='config.yml')
parser.add_argument(
'-v',
'--verbosity',
default='INFO',
choices=['DEBUG', 'ERROR', 'FATAL', 'INFO', 'WARM'],
)
args = parser.parse_args()
tf.logging.set_verbosity(args.verbosity)
params = utils.yaml_to_dict(args.config)
tf.logging.info("Using parameters: {}".format(params))
train_model(params)
meta_class = file_dir_split[-2]
sub_class = file_dir_split[-1]
image_path = os.path.join(root, name)
if meta_class not in feature_dict:
feature_dict[meta_class] = dict()
if sub_class not in feature_dict[meta_class]:
feature_dict[meta_class][sub_class] = list()
# read the training image
image = cv2.imread(image_path)
# convert the image to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# extract haralick texture from the image
print(i, ':', name)
textures = extract_features(gray)
feature_dict[meta_class][sub_class].append(textures)
i += 1
save_features(params, feature_dict)
if __name__ == '__main__':
params = utils.yaml_to_dict('config.yml')
download_data(params)
extract_data(params)
generate_textural_features(params)