深度学习图像识别数据集：CIFAR10 数据集（文件夹形式的十分类）

import os import matplotlib.pyplot as plt import random from PIL import Image # https://blog.csdn.net/qq_44886601/article/details/135486707?spm=1001.2014.3001.5502 def main(path): classes = [i for i in os.listdir(path)] # ['cat', 'dog'] # 将所有图片按照 类别:路径 字典形式保存 images_path = [] # [{'cat': './data/train\\cat\\Baidu_0000.jpeg'}, {'cat': './data/train\\cat\\Baidu_0002.jpeg'}] for cla in classes: for i in os.listdir(os.path.join(path, cla)): dic = {} # 类别:图像路径 img_path = os.path.join(path, cla, i) dic[cla] = img_path # {'cat': './data/train\\cat\\Baidu_0000.jpeg'} images_path.append(dic) # 随机展示4张图像 plt.figure(figsize=(12, 8)) for i in range(4): r = random.randint(0, len(images_path) - 1) # 生成随机数 label, im_path = list(images_path[r].keys())[0], list(images_path[r].values())[0] # cat , ./data/train\cat\Baidu_0049.jpeg im = Image.open(im_path) plt.subplot(2, 2, i + 1) plt.title(label) plt.imshow(im) plt.savefig('show.png') # 保存图片 plt.show() if __name__ == '__main__': root = './cifar10/train' # 传入目录 main(path=root)