crnn训练样本整理（1）：样本重新编号-1vs1的标签与图片（python）

BG：crnn做文字识别，训练样本需将图片与该图片的标签放到一个目录下，为便于管理尽可能多的样本，遂对样本进行重新编号。
编号规则：
图片：10位数字+’.[‘jpg’, ‘png’, ‘jpeg’]’
标签：10位数字+’.txt’

1vs1crnn样本重新编号.py

import os
import shutil

# 获取当前目录
dir = os.getcwd()

# path为批量文件的文件夹的路径
path = dir + '/train1'

new_path = dir + '/n_train1'

file_names = os.listdir(path)

# 设置起始编号
count = 1

img_format = ['jpg', 'png', 'jpeg']

for file_name in file_names:

    if file_name.split('.')[1] in img_format:

        print(file_name)
        img_name = file_name
        label_name = file_name.split('.')[0] + '.txt'

        new_img_name = '0'*(10-len(str(count))) + str(count) + '.' + file_name.split('.')[1]
        new_label_name = '0'*(10-len(str(count))) + str(count) + '.txt'

        img_path = os.path.join(path, file_name)
        label_path = os.path.join(path, label_name)

        new_img_path = os.path.join(new_path, new_img_name)
        new_label_path = os.path.join(new_path, new_label_name)

        try:
            shutil.move(img_path, new_img_path)
            shutil.move(label_path, new_label_path)
            count += 1
        except:
            pass

print('总计：', count-1)