model.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# The MIT License (MIT)
# Copyright (c) 2016 satojkovic

# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:

# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.

# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

import tensorflow as tf
import common

PATCH_SIZE = 5
NUM_CLASSES = len(common.CLASS_NAME)


def params():
    # weights and biases
    params = {}

    params['w_conv1'] = tf.get_variable(
        'w_conv1',
        shape=[PATCH_SIZE, PATCH_SIZE, common.CNN_IN_CH, 32],
        initializer=tf.contrib.layers.xavier_initializer())
    params['b_conv1'] = tf.Variable(tf.constant(0.1, shape=[32]))

    params['w_conv2'] = tf.get_variable(
        'w_conv2',
        shape=[PATCH_SIZE, PATCH_SIZE, 32, 64],
        initializer=tf.contrib.layers.xavier_initializer())
    params['b_conv2'] = tf.Variable(tf.constant(0.1, shape=[64]))

    params['w_conv3'] = tf.get_variable(
        'w_conv3',
        shape=[PATCH_SIZE, PATCH_SIZE, 64, 128],
        initializer=tf.contrib.layers.xavier_initializer())
    params['b_conv3'] = tf.Variable(tf.constant(0.1, shape=[128]))

    params['w_fc1'] = tf.get_variable(
        'w_fc1',
        shape=[16 * 4 * 128, 2048],
        initializer=tf.contrib.layers.xavier_initializer())
    params['b_fc1'] = tf.Variable(tf.constant(0.1, shape=[2048]))

    params['w_fc2'] = tf.get_variable(
        'w_fc2',
        shape=[2048, NUM_CLASSES],
        initializer=tf.contrib.layers.xavier_initializer())
    params['b_fc2'] = tf.Variable(tf.constant(0.1, shape=[NUM_CLASSES]))

    return params


def cnn(data, model_params, keep_prob):
    # First layer
    h_conv1 = tf.nn.relu(
        tf.nn.conv2d(
            data, model_params['w_conv1'], [1, 1, 1, 1], padding='SAME') +
        model_params['b_conv1'])
    h_pool1 = tf.nn.max_pool(
        h_conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')

    # Second layer
    h_conv2 = tf.nn.relu(
        tf.nn.conv2d(
            h_pool1, model_params['w_conv2'], [1, 1, 1, 1], padding='SAME') +
        model_params['b_conv2'])
    h_pool2 = tf.nn.max_pool(
        h_conv2, ksize=[1, 1, 2, 1], strides=[1, 1, 2, 1], padding='SAME')

    # Third layer
    h_conv3 = tf.nn.relu(
        tf.nn.conv2d(
            h_pool2, model_params['w_conv3'], [1, 1, 1, 1], padding='SAME') +
        model_params['b_conv3'])
    h_pool3 = tf.nn.max_pool(
        h_conv3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')

    # Fully connected layer
    conv_layer_flat = tf.reshape(h_pool3, [-1, 16 * 4 * 128])
    h_fc1 = tf.nn.relu(
        tf.matmul(conv_layer_flat, model_params['w_fc1']) +
        model_params['b_fc1'])
    h_fc1 = tf.nn.dropout(h_fc1, keep_prob)

    # Output layer
    out = tf.matmul(h_fc1, model_params['w_fc2']) + model_params['b_fc2']

    return out