Last active
January 23, 2018 08:27
-
-
Save sbugallo/aabd048bcb1d7ab26ece4c3499f826e0 to your computer and use it in GitHub Desktop.
Inception V3 for MRCNN
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# -*- coding: utf-8 -*- | |
from keras import layers | |
from keras.layers import Activation | |
from keras.layers import BatchNormalization | |
from keras.layers import Conv2D | |
from keras.layers import MaxPooling2D | |
from keras.layers import AveragePooling2D | |
from keras import backend as K | |
def conv2d_bn(x, | |
filters, | |
num_row, | |
num_col, | |
padding='same', | |
strides=(1, 1), | |
name=None): | |
""" | |
Utility function to apply conv + BN. | |
Arguments: | |
x: input tensor. | |
filters: filters in `Conv2D`. | |
num_row: height of the convolution kernel. | |
num_col: width of the convolution kernel. | |
padding: padding mode in `Conv2D`. | |
strides: strides in `Conv2D`. | |
name: name of the ops; will become `name + '_conv'` | |
for the convolution and `name + '_bn'` for the | |
batch norm layer. | |
Returns: | |
Output tensor after applying `Conv2D` and `BatchNormalization`. | |
""" | |
if name is not None: | |
bn_name = name + '_bn' | |
conv_name = name + '_conv' | |
else: | |
bn_name = None | |
conv_name = None | |
if K.image_data_format() == 'channels_first': | |
bn_axis = 1 | |
else: | |
bn_axis = 3 | |
x = Conv2D( | |
filters, (num_row, num_col), | |
strides=strides, | |
padding=padding, | |
use_bias=False, | |
name=conv_name)(x) | |
x = BatchNormalization(axis=bn_axis, scale=False, name=bn_name)(x) | |
x = Activation('relu', name=name)(x) | |
return x | |
def build_inception_graph(img_input, stage5=False): | |
if K.image_data_format() == 'channels_first': | |
channel_axis = 1 | |
else: | |
channel_axis = 3 | |
x = InceptionV3.conv2d_bn(img_input, 32, 3, 3, strides=(2, 2), padding='valid') | |
x = InceptionV3.conv2d_bn(x, 32, 3, 3, padding='valid') | |
x = InceptionV3.conv2d_bn(x, 64, 3, 3) | |
x = MaxPooling2D((3, 3), strides=(2, 2))(x) | |
x = InceptionV3.conv2d_bn(x, 80, 1, 1, padding='valid') | |
C2 = x = InceptionV3.conv2d_bn(x, 192, 3, 3, padding='valid') | |
x = MaxPooling2D((3, 3), strides=(2, 2))(x) | |
# mixed 0: 35 x 35 x 256 | |
branch1x1 = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(x, 48, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(branch5x5, 64, 5, 5) | |
branch3x3dbl = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 32, 1, 1) | |
x = layers.concatenate( | |
[branch1x1, branch5x5, branch3x3dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed0') | |
# mixed 1: 35 x 35 x 256 | |
branch1x1 = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(x, 48, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(branch5x5, 64, 5, 5) | |
branch3x3dbl = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 64, 1, 1) | |
x = layers.concatenate( | |
[branch1x1, branch5x5, branch3x3dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed1') | |
# mixed 2: 35 x 35 x 256 | |
branch1x1 = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(x, 48, 1, 1) | |
branch5x5 = InceptionV3.conv2d_bn(branch5x5, 64, 5, 5) | |
branch3x3dbl = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 64, 1, 1) | |
C3 = x = layers.concatenate( | |
[branch1x1, branch5x5, branch3x3dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed2') | |
# mixed 3: 17 x 17 x 768 | |
branch3x3 = InceptionV3.conv2d_bn(x, 384, 3, 3, strides=(2, 2), padding='valid') | |
branch3x3dbl = InceptionV3.conv2d_bn(x, 64, 1, 1) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 96, 3, 3, strides=(2, 2), padding='valid') | |
branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x) | |
x = layers.concatenate([branch3x3, branch3x3dbl, branch_pool], axis=channel_axis, name='mixed3') | |
# mixed 4: 17 x 17 x 768 | |
branch1x1 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(x, 128, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 128, 1, 7) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 192, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(x, 128, 1, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 128, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 128, 1, 7) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 128, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 1, 7) | |
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 192, 1, 1) | |
x = layers.concatenate( | |
[branch1x1, branch7x7, branch7x7dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed4') | |
# mixed 5, 6: 17 x 17 x 768 | |
for i in range(2): | |
branch1x1 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(x, 160, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 160, 1, 7) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 192, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(x, 160, 1, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 160, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 160, 1, 7) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 160, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 1, 7) | |
branch_pool = AveragePooling2D( | |
(3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 192, 1, 1) | |
x = layers.concatenate( | |
[branch1x1, branch7x7, branch7x7dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed' + str(5 + i)) | |
# mixed 7: 17 x 17 x 768 | |
branch1x1 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 192, 1, 7) | |
branch7x7 = InceptionV3.conv2d_bn(branch7x7, 192, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 1, 7) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 7, 1) | |
branch7x7dbl = InceptionV3.conv2d_bn(branch7x7dbl, 192, 1, 7) | |
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 192, 1, 1) | |
C4 = x = layers.concatenate( | |
[branch1x1, branch7x7, branch7x7dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed7') | |
# mixed 8: 8 x 8 x 1280 | |
branch3x3 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch3x3 = InceptionV3.conv2d_bn(branch3x3, 320, 3, 3, | |
strides=(2, 2), padding='valid') | |
branch7x7x3 = InceptionV3.conv2d_bn(x, 192, 1, 1) | |
branch7x7x3 = InceptionV3.conv2d_bn(branch7x7x3, 192, 1, 7) | |
branch7x7x3 = InceptionV3.conv2d_bn(branch7x7x3, 192, 7, 1) | |
branch7x7x3 = InceptionV3.conv2d_bn( | |
branch7x7x3, 192, 3, 3, strides=(2, 2), padding='valid') | |
branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x) | |
x = layers.concatenate( | |
[branch3x3, branch7x7x3, branch_pool], axis=channel_axis, name='mixed8') | |
# mixed 9, 10: 8 x 8 x 2048 | |
for i in range(2): | |
branch1x1 = InceptionV3.conv2d_bn(x, 320, 1, 1) | |
branch3x3 = InceptionV3.conv2d_bn(x, 384, 1, 1) | |
branch3x3_1 = InceptionV3.conv2d_bn(branch3x3, 384, 1, 3) | |
branch3x3_2 = InceptionV3.conv2d_bn(branch3x3, 384, 3, 1) | |
branch3x3 = layers.concatenate( | |
[branch3x3_1, branch3x3_2], axis=channel_axis, name='mixed9_' + str(i)) | |
branch3x3dbl = InceptionV3.conv2d_bn(x, 448, 1, 1) | |
branch3x3dbl = InceptionV3.conv2d_bn(branch3x3dbl, 384, 3, 3) | |
branch3x3dbl_1 = InceptionV3.conv2d_bn(branch3x3dbl, 384, 1, 3) | |
branch3x3dbl_2 = InceptionV3.conv2d_bn(branch3x3dbl, 384, 3, 1) | |
branch3x3dbl = layers.concatenate( | |
[branch3x3dbl_1, branch3x3dbl_2], axis=channel_axis) | |
branch_pool = AveragePooling2D( | |
(3, 3), strides=(1, 1), padding='same')(x) | |
branch_pool = InceptionV3.conv2d_bn(branch_pool, 192, 1, 1) | |
x = layers.concatenate( | |
[branch1x1, branch3x3, branch3x3dbl, branch_pool], | |
axis=channel_axis, | |
name='mixed' + str(9 + i)) | |
C5 = x if stage5 else None | |
return [None, C2, C3, C4, C5] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment