标注数据集数量不一致

最新推荐文章于 2022-10-29 21:51:21 发布

程序员阿明

最新推荐文章于 2022-10-29 21:51:21 发布

阅读量325

点赞数 1

文章标签：数据清理文件同步标注文件图像处理路径遍历

本文链接：https://blog.csdn.net/weixin_41463944/article/details/110849700

版权

import os
# 保存不同模型的目录名(绝对路径)
file_label = r'F:/抽烟/标注/'
file_picture = r'F:/抽烟/原图/'
#file_picture为原图存放路径，file_label为标注文件存放路径
def deleteTrainData(file_picture,file_label):
    label=[]
    picture=[]
    for root,dirs,files in os.walk(file_label):
        pass
    for root1,dirs1,filelabel in os.walk(file_picture):
        pass
    for i in range(len(files)):
        label.append(files[i][0:-4])
    for i in range(len(filelabel)):
        picture.append(filelabel[i][0:-4])
    for i in range(len(label)):
        if label[i] not in picture:
            delxmlPath = file_label + label[i] + ".xml"
            os.remove(delxmlPath)
            #label.remove(label[i])
    for i in range(len(picture)):
        if picture[i] not in label:
            delpicPath=file_picture+picture[i]+".jpg"
            os.remove(delpicPath)