Mmclassification annotation file generation

1 Data set composition

        mmclassification The dataset mainly contains four folders meta、train、val、test, among meta Stored annotation information , contain train.txt、val.txt and test.txt Three files .train、val、test Images of different categories are stored . Its composition is shown in the figure below ：

         train.txt、val.txt and test.txt The file stores the image path and category id, As shown in the figure below ：

        train、val、test Images of different categories are stored , Images of the same category are located in the same subfolder , The name of the subfolder is the corresponding category name , As shown in the figure below ：

         The data in the above example image comes from minist Handwritten font visualization dataset , According to train、test Folder for storage , Download at ：minist Handwritten digital visualization data set - Deep learning document resources -CSDN download .

2 Annotation file generation

         Annotation file generation refers to the generation without meta In the case of folders , according to train、val、test The folder is automatically generated meta Folder , And its folder train.txt、val.txt and test.txt.

3 Reference procedure

# -*- coding: utf-8 -*-
"""
 The official account of Lele perception school 
@author: https://blog.csdn.net/suiyingy
"""

import os
from glob import glob
from pathlib import Path

def generate_mmcls_ann(data_dir, img_type='.png'):
    data_dir = str(Path(data_dir)) + '/'
    classes = ['0', '1', '2', '3', '4','5', '6', '7', '8', '9'] 
    class2id = dict(zip(classes, range(len(classes))))
    data_dir  = str(Path(data_dir)) + '/'
    dir_types = ['train', 'val', 'test']
    sub_dirs = os.listdir(data_dir)
    ann_dir   = data_dir + 'meta/'
    if not os.path.exists(ann_dir):
        os.makedirs(ann_dir)
    for sd in sub_dirs:
        if sd not in dir_types:
            continue
        annotations = []
        target_dir  = data_dir + sd + '/'
        for d in os.listdir(target_dir):
            class_id = str(class2id[d])
            images = glob(target_dir+d+'/*'+img_type)
            for img in images:
                img = d + '/' + os.path.basename(img)
                annotations.append(img+' '+ class_id+'\n')
        annotations[-1] = annotations[-1].strip()
        with open(ann_dir+sd+'.txt', 'w') as f:
            f.writelines(annotations)


if __name__ == '__main__':
    data_dir = 'data/Minist Handwritten digital visualization data set /'
    generate_mmcls_ann(data_dir) 

4 Running results

         After running the above reference program ,data_dir The folder will automatically generate meta Folder and corresponding txt file ,txt The content of the document is shown in the figure below .

5 【python Three dimensional deep learning 】python 3D point cloud from basic to deep learning _Coding Leaves blog -CSDN Blog _python Three dimensional point cloud

More 3D 、 Please pay attention to two-dimensional perception algorithm and financial quantitative analysis algorithm “ Lele perception school ” WeChat official account , And will continue to update .