diff options
| -rw-r--r-- | README.md | 1 | ||||
| -rw-r--r-- | neural_style.py | 79 |
2 files changed, 40 insertions, 40 deletions
@@ -253,7 +253,6 @@ python neural_style.py --video \ * `--style_layers`: *Space-separated* VGG19 layer names used for the style image. *Default*: `relu1_1 relu2_1 relu3_1 relu4_1 relu5_1` * `--content_layer_weights`: Space-separated weights of each content layer to the content loss. *Default*: `1.0` * `--style_layer_weights`: Space-separated weights of each style layer to loss. *Default*: `0.2 0.2 0.2 0.2 0.2` -* `--style_scale`: Scale of the style image. Not currently implemented. * `--original_colors`: Boolean flag indicating if the style is transferred but not the colors. * `--style_mask`: Boolean flag indicating if style is transferred to masked regions. * `--style_mask_imgs`: Filenames of the style mask images (example: `face_mask.png`). To use multiple style mask images, pass a *space-separated* list. *Example*: `--style_mask_imgs face_mask.png face_mask_inv.png` diff --git a/neural_style.py b/neural_style.py index 565dbc0..374b4c1 100644 --- a/neural_style.py +++ b/neural_style.py @@ -4,8 +4,7 @@ import scipy.io import argparse import struct import time -import cv2 -import csv +import cv2 import os ''' @@ -61,7 +60,7 @@ def parse_args(): help='Weight for the style loss function. (default: %(default)s)') parser.add_argument('--tv_weight', type=float, - default=0, + default=1e-3, help='Weight for the transvariational loss function. Set small (e.g. 1e-3). (default: %(default)s)') parser.add_argument('--temporal_weight', type=float, @@ -88,9 +87,7 @@ def parse_args(): parser.add_argument('--style_layer_weights', nargs='+', type=float, default=[0.2, 0.2, 0.2, 0.2, 0.2], help='Contributions (weights) of each style layer to loss. (default: %(default)s)') - - parser.add_argument('--style_scale', type=float, default=1.0) - + parser.add_argument('--original_colors', action='store_true', help='Transfer the style but not the colors.') @@ -345,17 +342,12 @@ def content_layer_loss(p, x): if args.content_loss_function == 1: K = 1. / (2 * N**0.5 * M**0.5) elif args.content_loss_function == 2: - K = 1. / 2. + K = 1. / (N * M) elif args.content_loss_function == 3: - K = 1. / (N * M) + K = 1. / 2. loss = K * tf.reduce_sum(tf.pow((x - p), 2)) return loss -def gram_matrix(x, area, depth): - F = tf.reshape(x[0], (area, depth)) - G = tf.matmul(tf.transpose(F), F) - return G - def style_layer_loss(a, x): _, h, w, d = a.get_shape() M = h.value * w.value @@ -365,6 +357,11 @@ def style_layer_loss(a, x): loss = (1./(4 * N**2 * M**2)) * tf.reduce_sum(tf.pow((G - A), 2)) return loss +def gram_matrix(x, area, depth): + F = tf.reshape(x[0], (area, depth)) + G = tf.matmul(tf.transpose(F), F) + return G + def mask_style_layer(a, x, mask_img): _, h, w, d = a.get_shape() mask = get_mask_image(mask_img, w.value, h.value) @@ -443,7 +440,8 @@ def get_longterm_weights(i, j): c_max = tf.maximum(c - c_sum, 0.) return c_max -def sum_longterm_temporal_losses(net, frame, x): +def sum_longterm_temporal_losses(sess, net, frame, input_img): + x = sess.run(net['input'].assign(input_img)) loss = 0. for j in range(args.prev_frame_indices): prev_frame = frame - j @@ -452,7 +450,8 @@ def sum_longterm_temporal_losses(net, frame, x): loss += temporal_loss(x, w, c) return loss -def sum_shortterm_temporal_losses(net, frame, x): +def sum_shortterm_temporal_losses(sess, net, frame, input_img): + x = sess.run(net['input'].assign(input_img)) prev_frame = frame - 1 w = get_prev_warped_frame(frame) c = get_content_weights(frame, prev_frame) @@ -462,10 +461,11 @@ def sum_shortterm_temporal_losses(net, frame, x): ''' denoising loss function - remark: not convinced this does anything significant. + remark: not sure this does anything significant. ''' -def sum_total_variation_losses(x): - b, h, w, d = x.shape +def sum_total_variation_losses(sess, net, input_img): + b, h, w, d = input_img.shape + x = sess.run(net['input'].assign(input_img)) tv_y_size = b * (h-1) * w * d tv_x_size = b * h * (w-1) * d loss_y = tf.nn.l2_loss(x[:,1:,:,:] - x[:,:h-1,:,:]) @@ -480,7 +480,7 @@ def sum_total_variation_losses(x): utilities and i/o ''' def read_image(path): - # BGR image + # bgr image img = cv2.imread(path, cv2.IMREAD_COLOR).astype('float') img = preprocess(img, vgg19_mean) return img @@ -490,21 +490,19 @@ def write_image(path, img): cv2.imwrite(path, img) def preprocess(img, mean): - # BGR to RGB + # bgr to rgb img = img[...,::-1] # shape (h, w, d) to (1, h, w, d) img = img[np.newaxis,:,:,:] - # subtract mean img -= mean return img def postprocess(img, mean): - # add mean img += mean # shape (1, h, w, d) to (h, w, d) img = img[0] img = np.clip(img, 0, 255).astype('uint8') - # RGB to BGR + # rgb to bgr img = img[...,::-1] return img @@ -561,7 +559,7 @@ def stylize(content_img, style_imgs, init_img, frame=None): L_content = sum_content_losses(sess, net, content_img) # denoising loss - L_tv = sum_total_variation_losses(init_img) + L_tv = sum_total_variation_losses(sess, net, init_img) # loss weights alpha = args.content_weight @@ -575,7 +573,7 @@ def stylize(content_img, style_imgs, init_img, frame=None): if args.video and frame > 1: gamma = args.temporal_weight - L_temporal = sum_shortterm_temporal_losses(sess, frame, init_img) + L_temporal = sum_shortterm_temporal_losses(sess, net, frame, init_img) L_total += gamma * L_temporal # optimization algorithm @@ -589,7 +587,7 @@ def stylize(content_img, style_imgs, init_img, frame=None): output_img = sess.run(net['input']) if args.original_colors: - output_img = convert_to_original_colors(np.copy(content_img), np.copy(output_img)) + output_img = convert_to_original_colors(np.copy(content_img), output_img) if args.video: write_video_output(frame, output_img) @@ -597,14 +595,14 @@ def stylize(content_img, style_imgs, init_img, frame=None): write_image_output(output_img, content_img, style_imgs, init_img) def minimize_with_lbfgs(sess, net, optimizer, init_img): - if args.verbose: print('MINIMIZING LOSS USING: L-BFGS OPTIMIZER') + if args.verbose: print('\nMINIMIZING LOSS USING: L-BFGS OPTIMIZER') init_op = tf.initialize_all_variables() sess.run(init_op) sess.run(net['input'].assign(init_img)) optimizer.minimize(sess) def minimize_with_adam(sess, net, optimizer, init_img, loss): - if args.verbose: print('MINIMIZING LOSS USING: ADAM OPTIMIZER') + if args.verbose: print('\nMINIMIZING LOSS USING: ADAM OPTIMIZER') train_op = optimizer.minimize(loss) init_op = tf.initialize_all_variables() sess.run(init_op) @@ -612,6 +610,9 @@ def minimize_with_adam(sess, net, optimizer, init_img, loss): iterations = 0 while (iterations < args.max_iterations): sess.run(train_op) + if iterations % args.print_iterations == 0 and args.verbose: + curr_loss = loss.eval() + print("At iterate {}\tf= {:.5E}".format(iterations, curr_loss)) iterations += 1 def get_optimizer(loss): @@ -650,16 +651,16 @@ def write_image_output(output_img, content_img, style_imgs, init_img): # save the configuration settings out_file = os.path.join(out_dir, 'meta_data.txt') f = open(out_file, 'w') - f.write('image name: {}\n'.format(args.img_name)) + f.write('image_name: {}\n'.format(args.img_name)) f.write('content: {}\n'.format(args.content_img)) index = 0 for style_img, weight in zip(args.style_imgs, args.style_imgs_weights): - f.write('styles ['+str(index)+']: {} * {}\n'.format(weight, style_img)) + f.write('styles['+str(index)+']: {} * {}\n'.format(weight, style_img)) index += 1 index = 0 if args.style_mask_imgs is not None: for mask in args.style_mask_imgs: - f.write('style masks ['+str(index)+']: {}\n'.format(mask)) + f.write('style_masks['+str(index)+']: {}\n'.format(mask)) index += 1 f.write('init_type: {}\n'.format(args.init_img_type)) f.write('content_weight: {}\n'.format(args.content_weight)) @@ -698,8 +699,8 @@ def get_content_frame(frame): return img def get_content_image(content_img): - # BGR image path = os.path.join(args.content_img_dir, content_img) + # bgr image img = cv2.imread(path, cv2.IMREAD_COLOR).astype('float') h, w, d = img.shape mx = args.max_size @@ -713,14 +714,14 @@ def get_content_image(content_img): img = preprocess(img, vgg19_mean) return img -def get_style_images(content_img, scale): +def get_style_images(content_img): + _, ch, cw, cd = content_img.shape style_imgs = [] for style_fn in args.style_imgs: path = os.path.join(args.style_imgs_dir, style_fn) - # BGR image + # bgr image img = cv2.imread(path, cv2.IMREAD_COLOR).astype(np.float32) - _, h, w, d = content_img.shape - img = cv2.resize(img, dsize=(int(w*scale), int(h*scale))) + img = cv2.resize(img, dsize=(cw, ch)) img = preprocess(img, vgg19_mean) style_imgs.append(img) return style_imgs @@ -796,7 +797,7 @@ def convert_to_original_colors(content_img, stylized_img): def render_single_image(): content_img = get_content_image(args.content_img) - style_imgs = get_style_images(content_img, args.style_scale) + style_imgs = get_style_images(content_img) with tf.Graph().as_default(): print('\n---- RENDERING SINGLE IMAGE ----\n') init_img = get_init_image(args.init_img_type, content_img, style_imgs) @@ -811,7 +812,7 @@ def render_video(): print('\n---- RENDERING VIDEO FRAME: {}/{} ----\n'.format(frame, args.end_frame)) if frame == 1: content_frame = get_content_frame(frame) - style_imgs = get_style_images(content_frame, args.style_scale) + style_imgs = get_style_images(content_frame) init_img = get_init_image(args.first_frame_type, content_frame, style_imgs, frame) args.max_iterations = args.first_frame_iterations tick = time.time() @@ -820,7 +821,7 @@ def render_video(): print('Frame {} elapsed time: {}'.format(frame, tock - tick)) else: content_frame = get_content_frame(frame) - style_imgs = get_style_images(content_frame, args.style_scale) + style_imgs = get_style_images(content_frame) init_img = get_init_image(args.init_frame_type, content_frame, style_imgs, frame) args.max_iterations = args.frame_iterations tick = time.time() |