admin

Art Generation With Neural Style Transfer
Use transfer learning to combine a content image with a s...
扫描右侧二维码阅读全文
10
2018/07

Art Generation With Neural Style Transfer

Use transfer learning to combine a content image with a style image

import os
import sys
import scipy.io
import scipy.misc
import matplotlib.pyplot as plt
from matplotlib.pyplot import imshow
from PIL import Image
from nst_utils import *
import numpy as np
import tensorflow as tf

%matplotlib inline

model = load_vgg_model("pretrained-model/imagenet-vgg-verydeep-19.mat")
print(model)

content_image = scipy.misc.imread("images/louvre.jpg")
imshow(content_image)

def compute_content_cost(a_C, a_G):
    """
    Computes the content cost
    
    Arguments:
    a_C -- tensor of dimension (1, n_H, n_W, n_C), hidden layer activations representing content of the image C 
    a_G -- tensor of dimension (1, n_H, n_W, n_C), hidden layer activations representing content of the image G
    
    Returns: 
    J_content -- scalar that you compute using equation 1 above.
    """
    
    ### START CODE HERE ###
    # Retrieve dimensions from a_G (≈1 line)
    m, n_H, n_W, n_C = a_G.get_shape().as_list()
    
    # Reshape a_C and a_G (≈2 lines)
    # unroll the two tensors 
    a_C_unrolled = tf.reshape(a_C,[m,-1,n_C])
    a_G_unrolled = tf.reshape(a_G,[m,-1,n_C])
    #print(a_G_unrolled.get_shape().as_list())
    
    # compute the cost with tensorflow (≈1 line)
    J_content = tf.reduce_sum(tf.square(a_C_unrolled - a_G_unrolled))/(4*n_H*n_W*n_C)
    ### END CODE HERE ###
    
    return J_content

style_image = scipy.misc.imread("images/monet_800600.jpg")
imshow(style_image)

def gram_matrix(A):
    """
    Argument:
    A -- matrix of shape (n_C, n_H*n_W)
    
    Returns:
    GA -- Gram matrix of A, of shape (n_C, n_C)
    """
    
    ### START CODE HERE ### (≈1 line)
    GA = tf.matmul(A,tf.transpose(A))
    ### END CODE HERE ###
    
    return GA

def compute_layer_style_cost(a_S, a_G):
    """
    Arguments:
    a_S -- tensor of dimension (1, n_H, n_W, n_C), hidden layer activations representing style of the image S 
    a_G -- tensor of dimension (1, n_H, n_W, n_C), hidden layer activations representing style of the image G
    
    Returns: 
    J_style_layer -- tensor representing a scalar value, style cost defined above by equation (2)
    """
    
    ### START CODE HERE ###
    # Retrieve dimensions from a_G (≈1 line)
    m, n_H, n_W, n_C = a_G.get_shape().as_list()
    
    # Reshape the images to have them of shape (n_C, n_H*n_W) (≈2 lines)
    #a_S = tf.reshape(a_S,[n_C,-1])
    #a_G = tf.reshape(a_G,[n_C,-1])
    a_S = tf.transpose(tf.reshape(a_S,[n_H*n_W,n_C]))
    a_G = tf.transpose(tf.reshape(a_G,[n_H*n_W,n_C]))
    #print(a_S.get_shape().as_list())
    #print(a_G.get_shape().as_list())
    
    # Computing gram_matrices for both images S and G (≈2 lines)
    GS = gram_matrix(a_S)
    GG = gram_matrix(a_G)

    # Computing the loss (≈1 line)
    J_style_layer = (tf.reduce_sum(tf.square(tf.subtract(GS,GG))))/(4*n_C*n_C*(n_H*n_W)*(n_H*n_W))
    
    ### END CODE HERE ###
    
    return J_style_layer

STYLE_LAYERS = [
    ('conv1_1', 0.2),
    ('conv2_1', 0.2),
    ('conv3_1', 0.2),
    ('conv4_1', 0.2),
    ('conv5_1', 0.2)]

def compute_style_cost(model, STYLE_LAYERS):
    """
    Computes the overall style cost from several chosen layers
    
    Arguments:
    model -- our tensorflow model
    STYLE_LAYERS -- A python list containing:
                        - the names of the layers we would like to extract style from
                        - a coefficient for each of them
    
    Returns: 
    J_style -- tensor representing a scalar value, style cost defined above by equation (2)
    """
    
    # initialize the overall style cost
    J_style = 0

    for layer_name, coeff in STYLE_LAYERS:

        # Select the output tensor of the currently selected layer
        out = model[layer_name]

        # Set a_S to be the hidden layer activation from the layer we have selected, by running the session on out
        a_S = sess.run(out)

        # Set a_G to be the hidden layer activation from same layer. Here, a_G references model[layer_name] 
        # and isn't evaluated yet. Later in the code, we'll assign the image G as the model input, so that
        # when we run the session, this will be the activations drawn from the appropriate layer, with G as input.
        a_G = out
        
        # Compute style_cost for the current layer
        J_style_layer = compute_layer_style_cost(a_S, a_G)

        # Add coeff * J_style_layer of this layer to overall style cost
        J_style += coeff * J_style_layer

    return J_style

def total_cost(J_content, J_style, alpha = 10, beta = 40):
    """
    Computes the total cost function
    
    Arguments:
    J_content -- content cost coded above
    J_style -- style cost coded above
    alpha -- hyperparameter weighting the importance of the content cost
    beta -- hyperparameter weighting the importance of the style cost
    
    Returns:
    J -- total cost as defined by the formula above.
    """
    
    ### START CODE HERE ### (≈1 line)
    J = alpha*J_content + beta * J_style
    ### END CODE HERE ###
    
    return J

# Reset the graph
tf.reset_default_graph()

# Start interactive session
sess = tf.InteractiveSession()

content_image = scipy.misc.imread("images/louvre_small.jpg")
content_image = reshape_and_normalize_image(content_image)

style_image = scipy.misc.imread("images/monet.jpg")
style_image = reshape_and_normalize_image(style_image)

generated_image = generate_noise_image(content_image)
imshow(generated_image[0])

model = load_vgg_model("pretrained-model/imagenet-vgg-verydeep-19.mat")

# Assign the content image to be the input of the VGG model.  
sess.run(model['input'].assign(content_image))

# Select the output tensor of layer conv4_2
out = model['conv4_2']

# Set a_C to be the hidden layer activation from the layer we have selected
a_C = sess.run(out)

# Set a_G to be the hidden layer activation from same layer. Here, a_G references model['conv4_2'] 
# and isn't evaluated yet. Later in the code, we'll assign the image G as the model input, so that
# when we run the session, this will be the activations drawn from the appropriate layer, with G as input.
a_G = out
print(a_G)
print(a_C)

# Compute the content cost
J_content = compute_content_cost(a_C, a_G)

# Assign the input of the model to be the "style" image 
sess.run(model['input'].assign(style_image))

# Compute the style cost
J_style = compute_style_cost(model, STYLE_LAYERS)

J = total_cost(J_content,J_style)

# define optimizer (1 line)
optimizer = tf.train.AdamOptimizer(2.0)

# define train_step (1 line)
train_step = optimizer.minimize(J)

def model_nn(sess, input_image, num_iterations = 200):
    
    # Initialize global variables (you need to run the session on the initializer)
    ### START CODE HERE ### (1 line)
    sess.run(tf.global_variables_initializer())
    ### END CODE HERE ###
    
    # Run the noisy input image (initial generated image) through the model. Use assign().
    ### START CODE HERE ### (1 line)
    sess.run(model['input'].assign(input_image))
    ### END CODE HERE ###
    
    for i in range(num_iterations):
    
        # Run the session on the train_step to minimize the total cost
        ### START CODE HERE ### (1 line)
        sess.run(train_step)
        ### END CODE HERE ###
        
        # Compute the generated image by running the session on the current model['input']
        ### START CODE HERE ### (1 line)
        generated_image = sess.run(model['input'])
        ### END CODE HERE ###

        # Print every 20 iteration.
        if i%20 == 0:
            Jt, Jc, Js = sess.run([J, J_content, J_style])
            print("Iteration " + str(i) + " :")
            print("total cost = " + str(Jt))
            print("content cost = " + str(Jc))
            print("style cost = " + str(Js))
            
            # save current generated image in the "/output" directory
            save_image("output/" + str(i) + ".png", generated_image)
    
    # save last generated image
    save_image('output/generated_image.jpg', generated_image)
    
    return generated_image

model_nn(sess, generated_image)
  • The idea of using a network trained on a different task and applying it to a new task is called transfer learning. In this project we use the transfer learning and load the pre-trained VGG-19 network.
  • Pay attention to the function tf.add(),tf.subtract(),tf.square(), tf.reduce_sum(), tf.reshape(X,[2,2,-1]),tf.transpose(X, perm = [0,2,1]), you can look up them in the official API docs if necessary.
  • The loss consists of the J_content and the J_style, pay attention to the gram matrix and the computation of the J_style.
  • Understand the two unrollments of the tensors with tf.reshape() function and compare their differences. 
Last modification:March 13th, 2019 at 07:06 pm
If you think my article is useful to you, please feel free to appreciate

Leave a Comment