flowers.py 6.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This module will download dataset from
16
http://www.robots.ox.ac.uk/~vgg/data/flowers/102/index.html
17 18
and parse train/test set intopaddle reader creators.

19
This set contains images of flowers belonging to 102 different categories.
20 21 22 23 24 25
The images were acquired by searching the web and taking pictures. There are a
minimum of 40 images for each category.

The database was used in:

Nilsback, M-E. and Zisserman, A. Automated flower classification over a large
26 27
 number of classes.Proceedings of the Indian Conference on Computer Vision,
Graphics and Image Processing (2008)
28 29 30 31 32 33 34 35
http://www.robots.ox.ac.uk/~vgg/publications/papers/nilsback08.{pdf,ps.gz}.

"""
import cPickle
import itertools
from common import download
import tarfile
import scipy.io as scio
36
from paddle.v2.image import *
37
from paddle.v2.reader import *
38 39
import os
import numpy as np
40
from multiprocessing import cpu_count
41 42 43 44 45 46 47 48
__all__ = ['train', 'test', 'valid']

DATA_URL = 'http://www.robots.ox.ac.uk/~vgg/data/flowers/102/102flowers.tgz'
LABEL_URL = 'http://www.robots.ox.ac.uk/~vgg/data/flowers/102/imagelabels.mat'
SETID_URL = 'http://www.robots.ox.ac.uk/~vgg/data/flowers/102/setid.mat'
DATA_MD5 = '52808999861908f626f3c1f4e79d11fa'
LABEL_MD5 = 'e0620be6f572b9609742df49c70aed4d'
SETID_MD5 = 'a5357ecc9cb78c4bef273ce3793fc85c'
W
wanghaoshuang 已提交
49 50 51 52 53 54
# In official 'readme', tstid is the flag of test data
# and trnid is the flag of train data. But test data is more than train data.
# So we exchange the train data and test data.
TRAIN_FLAG = 'tstid'
TEST_FLAG = 'trnid'
VALID_FLAG = 'valid'
55 56 57 58 59 60 61


def default_mapper(sample):
    '''
    map image bytes data to type needed by model input layer
    '''
    img, label = sample
62 63
    img = load_image_bytes(img)
    img = simple_transform(img, 256, 224, True)
64 65 66 67 68 69
    return img.flatten().astype('float32'), label


def reader_creator(data_file,
                   label_file,
                   setid_file,
70 71
                   dataset_name,
                   mapper=default_mapper,
72
                   buffered_size=1024,
W
wanghaoshuang 已提交
73
                   use_xmap=True):
74
    '''
75
    1. read images from tar file and
76 77
        merge images into batch files in 102flowers.tgz_batch/
    2. get a reader to read sample from batch file
78 79

    :param data_file: downloaded data file
80
    :type data_file: string
81
    :param label_file: downloaded label file
82 83 84 85
    :type label_file: string
    :param setid_file: downloaded setid file containing information
                        about how to split dataset
    :type setid_file: string
86 87
    :param dataset_name: data set name (tstid|trnid|valid)
    :type dataset_name: string
88
    :param mapper: a function to map image bytes data to type
89 90
                    needed by model input layer
    :type mapper: callable
91 92
    :param buffered_size: the size of buffer used to process images
    :type buffered_size: int
93 94 95
    :return: data reader
    :rtype: callable
    '''
96 97 98 99 100 101 102
    labels = scio.loadmat(label_file)['labels'][0]
    indexes = scio.loadmat(setid_file)[dataset_name][0]
    img2label = {}
    for i in indexes:
        img = "jpg/image_%05d.jpg" % i
        img2label[img] = labels[i - 1]
    file_list = batch_images_from_tar(data_file, dataset_name, img2label)
103 104 105 106 107 108 109 110 111 112 113 114

    def reader():
        for file in open(file_list):
            file = file.strip()
            batch = None
            with open(file, 'r') as f:
                batch = cPickle.load(f)
            data = batch['data']
            labels = batch['label']
            for sample, label in itertools.izip(data, batch['label']):
                yield sample, int(label)

W
wanghaoshuang 已提交
115
    if use_xmap:
116 117 118
        return xmap_readers(mapper, reader, cpu_count(), buffered_size)
    else:
        return map_readers(mapper, reader)
119 120


W
wanghaoshuang 已提交
121
def train(mapper=default_mapper, buffered_size=1024, use_xmap=True):
122
    '''
123 124 125
    Create flowers training set reader.
    It returns a reader, each sample in the reader is
    image pixels in [0, 1] and label in [1, 102]
126 127 128 129 130 131
    translated from original color image by steps:
    1. resize to 256*256
    2. random crop to 224*224
    3. flatten
    :param mapper:  a function to map sample.
    :type mapper: callable
132 133
    :param buffered_size: the size of buffer used to process images
    :type buffered_size: int
134 135 136 137 138 139
    :return: train data reader
    :rtype: callable
    '''
    return reader_creator(
        download(DATA_URL, 'flowers', DATA_MD5),
        download(LABEL_URL, 'flowers', LABEL_MD5),
W
wanghaoshuang 已提交
140 141
        download(SETID_URL, 'flowers', SETID_MD5), TRAIN_FLAG, mapper,
        buffered_size, use_xmap)
142 143


W
wanghaoshuang 已提交
144
def test(mapper=default_mapper, buffered_size=1024, use_xmap=True):
145
    '''
146 147 148
    Create flowers test set reader.
    It returns a reader, each sample in the reader is
    image pixels in [0, 1] and label in [1, 102]
149 150 151 152 153 154
    translated from original color image by steps:
    1. resize to 256*256
    2. random crop to 224*224
    3. flatten
    :param mapper:  a function to map sample.
    :type mapper: callable
155 156
    :param buffered_size: the size of buffer used to process images
    :type buffered_size: int
157 158 159 160 161 162
    :return: test data reader
    :rtype: callable
    '''
    return reader_creator(
        download(DATA_URL, 'flowers', DATA_MD5),
        download(LABEL_URL, 'flowers', LABEL_MD5),
W
wanghaoshuang 已提交
163 164
        download(SETID_URL, 'flowers', SETID_MD5), TEST_FLAG, mapper,
        buffered_size, use_xmap)
165 166


W
wanghaoshuang 已提交
167
def valid(mapper=default_mapper, buffered_size=1024, use_xmap=True):
168
    '''
169 170 171
    Create flowers validation set reader.
    It returns a reader, each sample in the reader is
    image pixels in [0, 1] and label in [1, 102]
172 173 174 175
    translated from original color image by steps:
    1. resize to 256*256
    2. random crop to 224*224
    3. flatten
176 177 178 179 180 181
    :param mapper:  a function to map sample.
    :type mapper: callable
    :param buffered_size: the size of buffer used to process images
    :type buffered_size: int
    :return: test data reader
    :rtype: callable
182 183 184 185
    '''
    return reader_creator(
        download(DATA_URL, 'flowers', DATA_MD5),
        download(LABEL_URL, 'flowers', LABEL_MD5),
W
wanghaoshuang 已提交
186 187
        download(SETID_URL, 'flowers', SETID_MD5), VALID_FLAG, mapper,
        buffered_size, use_xmap)
188 189 190 191 192 193


def fetch():
    download(DATA_URL, 'flowers', DATA_MD5)
    download(LABEL_URL, 'flowers', LABEL_MD5)
    download(SETID_URL, 'flowers', SETID_MD5)