tiny_dnn 1.0.0
A header only, dependency-free deep learning framework in C++11
Loading...
Searching...
No Matches
mnist_parser.h
1/*
2 Copyright (c) 2013, Taiga Nomi
3 All rights reserved.
4
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions are met:
7 * Redistributions of source code must retain the above copyright
8 notice, this list of conditions and the following disclaimer.
9 * Redistributions in binary form must reproduce the above copyright
10 notice, this list of conditions and the following disclaimer in the
11 documentation and/or other materials provided with the distribution.
12 * Neither the name of the <organization> nor the
13 names of its contributors may be used to endorse or promote products
14 derived from this software without specific prior written permission.
15
16 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
17 EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
18 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
19 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY
20 DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
21 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
22 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
23 ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26*/
27#pragma once
28#include "tiny_dnn/util/util.h"
29#include <fstream>
30#include <cstdint>
31
32namespace tiny_dnn {
33namespace detail {
34
36 uint32_t magic_number;
37 uint32_t num_items;
38 uint32_t num_rows;
39 uint32_t num_cols;
40};
41
42inline void parse_mnist_header(std::ifstream& ifs, mnist_header& header) {
43 ifs.read((char*) &header.magic_number, 4);
44 ifs.read((char*) &header.num_items, 4);
45 ifs.read((char*) &header.num_rows, 4);
46 ifs.read((char*) &header.num_cols, 4);
47
48 if (is_little_endian()) {
49 reverse_endian(&header.magic_number);
50 reverse_endian(&header.num_items);
51 reverse_endian(&header.num_rows);
52 reverse_endian(&header.num_cols);
53 }
54
55 if (header.magic_number != 0x00000803 || header.num_items <= 0)
56 throw nn_error("MNIST label-file format error");
57 if (ifs.fail() || ifs.bad())
58 throw nn_error("file error");
59}
60
61inline void parse_mnist_image(std::ifstream& ifs,
62 const mnist_header& header,
63 float_t scale_min,
64 float_t scale_max,
65 int x_padding,
66 int y_padding,
67 vec_t& dst) {
68 const int width = header.num_cols + 2 * x_padding;
69 const int height = header.num_rows + 2 * y_padding;
70
71 std::vector<uint8_t> image_vec(header.num_rows * header.num_cols);
72
73 ifs.read((char*) &image_vec[0], header.num_rows * header.num_cols);
74
75 dst.resize(width * height, scale_min);
76
77 for (uint32_t y = 0; y < header.num_rows; y++)
78 for (uint32_t x = 0; x < header.num_cols; x++)
79 dst[width * (y + y_padding) + x + x_padding]
80 = (image_vec[y * header.num_cols + x] / float_t(255)) * (scale_max - scale_min) + scale_min;
81}
82
83} // namespace detail
84
92inline void parse_mnist_labels(const std::string& label_file, std::vector<label_t> *labels) {
93 std::ifstream ifs(label_file.c_str(), std::ios::in | std::ios::binary);
94
95 if (ifs.bad() || ifs.fail())
96 throw nn_error("failed to open file:" + label_file);
97
98 uint32_t magic_number, num_items;
99
100 ifs.read((char*) &magic_number, 4);
101 ifs.read((char*) &num_items, 4);
102
103 if (is_little_endian()) { // MNIST data is big-endian format
104 reverse_endian(&magic_number);
105 reverse_endian(&num_items);
106 }
107
108 if (magic_number != 0x00000801 || num_items <= 0)
109 throw nn_error("MNIST label-file format error");
110
111 for (uint32_t i = 0; i < num_items; i++) {
112 uint8_t label;
113 ifs.read((char*) &label, 1);
114 labels->push_back((label_t) label);
115 }
116}
117
140inline void parse_mnist_images(const std::string& image_file,
141 std::vector<vec_t> *images,
142 float_t scale_min,
143 float_t scale_max,
144 int x_padding,
145 int y_padding) {
146
147 if (x_padding < 0 || y_padding < 0)
148 throw nn_error("padding size must not be negative");
149 if (scale_min >= scale_max)
150 throw nn_error("scale_max must be greater than scale_min");
151
152 std::ifstream ifs(image_file.c_str(), std::ios::in | std::ios::binary);
153
154 if (ifs.bad() || ifs.fail())
155 throw nn_error("failed to open file:" + image_file);
156
157 detail::mnist_header header;
158
159 detail::parse_mnist_header(ifs, header);
160
161 for (uint32_t i = 0; i < header.num_items; i++) {
162 vec_t image;
163 detail::parse_mnist_image(ifs, header, scale_min, scale_max, x_padding, y_padding, image);
164 images->push_back(image);
165 }
166}
167
168} // namespace tiny_dnn
Simple image utility class.
Definition image.h:94
error exception class for tiny-dnn
Definition nn_error.h:37
Definition mnist_parser.h:35