generate_list.py

# -*- coding: utf-8 -*-
"""
========================================================================
A siamese-like CNN for diabetic retinopathy detection using binocular 
fudus images as input, Version 1.0
Copyright(c) 2020 Xianglong Zeng, Haiquan Chen, Yuan Luo, Wenbin Ye
All Rights Reserved.
----------------------------------------------------------------------
Permission to use, copy, or modify this software and its documentation
for educational and research purposes only and without fee is here
granted, provided that this copyright notice and the original authors'
names appear on all copies and supporting documentation. This program
shall not be used, rewritten, or adapted as the basis of a commercial
software or hardware product without first obtaining permission of the
authors. The authors make no representations about the suitability of
this software for any purpose. It is provided "as is" without express
or implied warranty.
----------------------------------------------------------------------
Please cite the following paper when you use it:
X. Zeng, H. Chen, Y. Luo and W. Ye, "Automated Diabetic Retinopathy 
Detection Based on Binocular Siamese-Like Convolutional Neural 
Network," in IEEE Access, vol. 7, pp. 30744-30753, 2019, 
doi: 10.1109/ACCESS.2019.2903171.
----------------------------------------------------------------------
This file use to generate the lists containing the image paths and 
labels of left and right eyes respectively, using for training and 
validation.

@author: Xianglong Zeng
========================================================================
"""

from sklearn.utils import shuffle

# 训练集/验证集划分比例
val_size = 0.2
# 图片文件夹路径
doc_path = '/shared_folders/train_data/'

csv_lines = open('./list/trainLabels_referable.csv').readlines()[1:]
print('# total images: {}'.format(len(csv_lines)))

train_lines_left = []
val_lines_left = []
train_lines_right = []
val_lines_right = []
dic_left_name = {}
dic_right_name = {}
label_00 = []
label_11 = []
label_01 = []
label_10 = []
dataset = []

for line in csv_lines:
    image_name, label = line.strip().split(',')
    # 按【标签：名字】构建左眼数据字典
    if image_name[-4] == 'l':
        dic_left_name[image_name] = label
    # 按【名字：标签】构建右眼数据字典
    if image_name[-5] == 'r':
        dic_right_name[image_name] = label

for left_name, right_name in zip(dic_left_name.keys(),dic_right_name.keys()):
    left_label = dic_left_name[left_name]
    right_label = dic_right_name[right_name]
    if left_label == right_label == '0':
        label_00.append(left_name[:-4])
    elif left_label == right_label == '1':
        label_11.append(left_name[:-4])
    elif left_label == '0':
        label_01.append(left_name[:-4])
    elif left_label == '1':
        label_10.append(left_name[:-4])

for data, label in zip([label_00,label_11,label_01,label_10], ['00','11','01','10']):
    data = shuffle(data)
    num_val = int(len(data) * val_size)
    print('eyes with type {} and {} respectively: {} pairs'.format(label[0], label[1], len(data)))
    for image_name in data[:num_val]:
        to_write_left = doc_path + '{}left.jpeg {}\n'.format(image_name, label[0])
        to_write_right = doc_path + '{}right.jpeg {}\n'.format(image_name, label[1])
        val_lines_left.append(to_write_left)
        val_lines_right.append(to_write_right)
    for image_name in data[num_val:]:
        to_write_left = doc_path + '{}left.jpeg {}\n'.format(image_name, label[0])
        to_write_right = doc_path + '{}right.jpeg {}\n'.format(image_name, label[1])
        train_lines_left.append(to_write_left)
        train_lines_right.append(to_write_right)

#将左右眼训练集按相同随机状态打乱
train_lines_left = shuffle(train_lines_left, random_state = 9527)
train_lines_right = shuffle(train_lines_right, random_state = 9527)
#将左右眼验证集按相同随机状态打乱
val_lines_left = shuffle(val_lines_left, random_state = 2333)
val_lines_right = shuffle(val_lines_right, random_state = 2333)

print('# train of left: {}, # train of right: {}\n# val of left: {}, # val of right: {}' \
      .format(len(train_lines_left), len(train_lines_right), len(val_lines_left), len(val_lines_right)))

with open('./list/kaggle_train_left.list', 'w') as f_train_left:
    for train_line in train_lines_left:
        f_train_left.write(train_line)

with open('./list/kaggle_train_right.list', 'w') as f_train_right:
    for train_line in train_lines_right:
        f_train_right.write(train_line)

with open('./list/kaggle_val_left.list', 'w') as f_val_left:
    for val_line in val_lines_left:
        f_val_left.write(val_line)

with open('./list/kaggle_val_right.list', 'w') as f_val_right:
    for val_line in val_lines_right:
        f_val_right.write(val_line)

print('kaggle_train.list and kaggle_val.list generated!')