parser = argparse.ArgumentParser('W2VVPP extract features script.')

parser.add_argument('--data_path', type=str,

help='image dataset path')

parser.add_argument('--feature_path', type=str,

help='specified path to file to save features')

args = parser.parse_args()

opt = parse_args()

print(json.dumps(vars(opt), indent = 2))

input_path = opt.data_path

print("Input path:"+str(input_path))

output_path = opt.feature_path

print("Output path:"+(output_path))

#loading pretrain resnet 152

resnet152_model = ResNet152(ResNet152(include_top=False, weights='imagenet', input_tensor=None, input_shape=None, pooling=None,))

count = 0

features_filename = open(output_path,"w")

for filename in os.listdir(input_path):

if filename.endswith(".jpg") or filename.endswith(".png"):

image_path = os.path.join(input_path,filename)

print("Image path:",image_path)

img = image.load_img(image_path, target_size=(224, 224))

x = image.img_to_array(img)

x = np.expand_dims(x, axis=0)

x = preprocess_input(x)

features = resnet152_model.predict(x)[0]

output = filename+" "+" ".join(np.asarray(features, dtype=str))+"\n"

# filename_features_format = np.insert(features.astype(str), 0, str(filename), axis=0)

features_filename.write(output)

# np.savetxt(features_filename,filename_features_format,fmt='%s')

count += 1