157 lines
7.2 KiB
Python
157 lines
7.2 KiB
Python
|
import cv2
|
||
|
|
import os
|
||
|
|
import tensorflow as tf
|
||
|
|
import numpy as np
|
||
|
|
from sklearn.model_selection import train_test_split
|
||
|
|
|
||
|
|
dir_img="CameraRGB/"
|
||
|
|
dir_mask="CameraSeg/"
|
||
|
|
|
||
|
|
width=200
|
||
|
|
height=125
|
||
|
|
|
||
|
|
taille_batch=100
|
||
|
|
nbr_entrainement=100
|
||
|
|
|
||
|
|
def crop(tensor1, tensor2):
|
||
|
|
offsets=(0, (int(tensor1.get_shape()[1])-int(tensor2.get_shape()[1]))//2, (int(tensor1.get_shape()[2])-int(tensor2.get_shape()[2]))//2, 0)
|
||
|
|
size=(-1, int(tensor2.get_shape()[1]), int(tensor2.get_shape()[2]), -1)
|
||
|
|
return tf.slice(tensor1, offsets, size)
|
||
|
|
|
||
|
|
def convolution(input, taille_noyau, nbr_cc, stride, b_norm=False, f_activation=None, training=False, padding='SAME'):
|
||
|
|
w=tf.Variable(tf.random.truncated_normal(shape=(taille_noyau, taille_noyau, int(input.get_shape()[-1]), nbr_cc)))
|
||
|
|
b=np.zeros(nbr_cc)
|
||
|
|
result=tf.nn.conv2d(input, w, strides=[1, stride, stride, 1], padding=padding)
|
||
|
|
result=tf.nn.bias_add(result, b)
|
||
|
|
if b_norm is True:
|
||
|
|
result=tf.layers.batch_normalization(result, training=training)
|
||
|
|
if f_activation is not None:
|
||
|
|
result=f_activation(result)
|
||
|
|
return result
|
||
|
|
|
||
|
|
def deconvolution(input, taille_noyau, nbr_cc, stride, b_norm=False, f_activation=None, training=False, padding='SAME'):
|
||
|
|
w=tf.Variable(tf.random.truncated_normal(shape=(taille_noyau, taille_noyau, nbr_cc, int(input.get_shape()[-1]))))
|
||
|
|
b=np.zeros(nbr_cc)
|
||
|
|
if padding == 'VALID':
|
||
|
|
out_h=(int(input.get_shape()[1])-1)*stride+taille_noyau
|
||
|
|
out_w=(int(input.get_shape()[2])-1)*stride+taille_noyau
|
||
|
|
elif padding == 'SAME':
|
||
|
|
out_h=(int(input.get_shape()[1])-1)*stride+1
|
||
|
|
out_w=(int(input.get_shape()[2])-1)*stride+1
|
||
|
|
else:
|
||
|
|
quit("erreur padding")
|
||
|
|
b_size=tf.shape(input)[0]
|
||
|
|
result=tf.nn.conv2d_transpose(input, w, output_shape=[b_size, out_h, out_w, nbr_cc], strides=[1, stride, stride, 1], padding=padding)
|
||
|
|
result=tf.nn.bias_add(result, b)
|
||
|
|
if b_norm is True:
|
||
|
|
result=tf.layers.batch_normalization(result, training=training)
|
||
|
|
if f_activation is not None:
|
||
|
|
result=f_activation(result)
|
||
|
|
return result
|
||
|
|
|
||
|
|
def unet(nbr_mask, size, padding='SAME', learning_rate=1E-3):
|
||
|
|
ph_images=tf.placeholder(shape=(None, size[0], size[1], size[2]), dtype=tf.float32, name='entree')
|
||
|
|
ph_masks=tf.placeholder(shape=(None, size[0], size[1], nbr_mask), dtype=tf.float32)
|
||
|
|
ph_is_training=tf.placeholder_with_default(False, (), name='is_training')
|
||
|
|
|
||
|
|
result=convolution(ph_images, 3, 32, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c1=convolution(result, 3, 32, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=tf.nn.max_pool(c1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
|
||
|
|
|
||
|
|
result=convolution(result, 3, 64, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c2=convolution(result, 3, 64, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=tf.nn.max_pool(c2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
|
||
|
|
|
||
|
|
result=convolution(result, 3, 128, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c3=convolution(result, 3, 128, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=tf.nn.max_pool(c3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
|
||
|
|
|
||
|
|
result=convolution(result, 3, 256, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=convolution(result, 3, 256, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
|
||
|
|
d3=deconvolution(result, 3, 256, 2, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c3=crop(c3, d3)
|
||
|
|
result=tf.concat((d3, c3), axis=3)
|
||
|
|
|
||
|
|
result=convolution(result, 3, 128, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=convolution(result, 3, 128, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
|
||
|
|
d2=deconvolution(result, 3, 128, 2, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c2=crop(c2, d2)
|
||
|
|
result=tf.concat((d2, c2), axis=3)
|
||
|
|
|
||
|
|
result=convolution(result, 3, 64, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=convolution(result, 3, 64, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
|
||
|
|
d1=deconvolution(result, 3, 64, 2, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
c1=crop(c1, d1)
|
||
|
|
result=tf.concat((d1, c1), axis=3)
|
||
|
|
|
||
|
|
result=convolution(result, 3, 32, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
result=convolution(result, 3, 32, 1, True, tf.nn.relu, ph_is_training, padding)
|
||
|
|
|
||
|
|
result=convolution(result, 1, nbr_mask, 1, False, None, ph_is_training, padding)
|
||
|
|
result=tf.image.resize_images(result, (ph_masks.get_shape()[1], ph_masks.get_shape()[2]))
|
||
|
|
mask=tf.nn.sigmoid(result, name="sortie")
|
||
|
|
|
||
|
|
loss=tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(labels=ph_masks, logits=result))
|
||
|
|
accuracy=tf.reduce_mean(tf.cast(tf.equal(tf.round(mask), ph_masks), tf.float32))
|
||
|
|
|
||
|
|
extra_update_ops=tf.get_collection(tf.GraphKeys.UPDATE_OPS)
|
||
|
|
with tf.control_dependencies(extra_update_ops):
|
||
|
|
train=tf.train.AdamOptimizer(learning_rate).minimize(loss)
|
||
|
|
|
||
|
|
return ph_images, ph_masks, ph_is_training, mask, train, accuracy, tf.train.Saver()
|
||
|
|
|
||
|
|
ph_images, ph_masks, ph_is_training, mask, train, accuracy, saver=unet(2, (height, width, 3), 'VALID')
|
||
|
|
|
||
|
|
tab_img=[]
|
||
|
|
tab_mask=[]
|
||
|
|
#for dir in ['../tutoriel19-1/dataA/', '../tutoriel19-1/dataB/', '../tutoriel19-1/dataC/', '../tutoriel19-1/dataD/', '../tutoriel19-1/dataE/']:
|
||
|
|
for dir in ['../tutoriel19-1/dataA/', '../tutoriel19-1/dataB/', '../tutoriel19-1/dataC/', '../tutoriel19-1/dataD/']:
|
||
|
|
for file in os.listdir(dir+dir_img):
|
||
|
|
tab_img.append(cv2.resize(cv2.imread(dir+dir_img+file)[0:500, 0:800], (width, height))/255)
|
||
|
|
img_mask=cv2.resize(cv2.imread(dir+dir_mask+file)[0:500, 0:800], (width, height))[:,:,2]
|
||
|
|
img_mask_result=np.zeros(shape=(height, width, 2), dtype=np.float32)
|
||
|
|
img_mask_result[:,:,0][img_mask==10]=1.
|
||
|
|
img_mask_result[:,:,1][img_mask==12]=1.
|
||
|
|
tab_mask.append(img_mask_result)
|
||
|
|
|
||
|
|
tab_img=np.array(tab_img)
|
||
|
|
tab_mask=np.array(tab_mask)
|
||
|
|
|
||
|
|
train_images, test_images, train_labels, test_labels=train_test_split(tab_img, tab_mask, test_size=.05)
|
||
|
|
|
||
|
|
with tf.Session() as s:
|
||
|
|
s.run(tf.global_variables_initializer())
|
||
|
|
tab_train=[]
|
||
|
|
tab_test=[]
|
||
|
|
for id_entrainement in np.arange(nbr_entrainement):
|
||
|
|
print("> Entrainement", id_entrainement)
|
||
|
|
for batch in np.arange(0, len(train_images), taille_batch):
|
||
|
|
s.run(train, feed_dict={
|
||
|
|
ph_images: train_images[batch:batch+taille_batch],
|
||
|
|
ph_masks: train_labels[batch:batch+taille_batch],
|
||
|
|
ph_is_training: True
|
||
|
|
})
|
||
|
|
print(" entrainement OK")
|
||
|
|
tab_accuracy_train=[]
|
||
|
|
for batch in np.arange(0, len(train_images), taille_batch):
|
||
|
|
p=s.run(accuracy, feed_dict={
|
||
|
|
ph_images: train_images[batch:batch+taille_batch],
|
||
|
|
ph_masks: train_labels[batch:batch+taille_batch]
|
||
|
|
})
|
||
|
|
tab_accuracy_train.append(p)
|
||
|
|
print(" train:", np.mean(tab_accuracy_train))
|
||
|
|
tab_accuracy_test=[]
|
||
|
|
for batch in np.arange(0, len(test_images), taille_batch):
|
||
|
|
p=s.run(accuracy, feed_dict={
|
||
|
|
ph_images: test_images[batch:batch+taille_batch],
|
||
|
|
ph_masks: test_labels[batch:batch+taille_batch]
|
||
|
|
})
|
||
|
|
tab_accuracy_test.append(p)
|
||
|
|
print(" test :", np.mean(tab_accuracy_test))
|
||
|
|
tab_train.append(1-np.mean(tab_accuracy_train))
|
||
|
|
tab_test.append(1-np.mean(tab_accuracy_test))
|
||
|
|
saver.save(s, './mon_modele/modele')
|