python将cifar-10二进制数据集转成图片格式的训练集和测试集

CIFAR-10数据集

CIFAR-10和CIFAR-100是来自于80 million张小型图片的数据集，图片收集者是Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton。
官网 http://www.cs.toronto.edu/~kriz/cifar.html

测试集和验证集效果：

CIFAR-10数据集

#encoding:utf-8
from cv2 import imwrite as imsave
import numpy as np
import os 

# 解压缩，返回解压后的字典
def unpickle(file):
    import pickle
    with open(file, 'rb') as fo:
        dicts = pickle.load(fo, encoding='bytes')
    return dicts
#创建文件夹
date_dict = {0:"airplane",1:"automobile",2:"bird",3:"cat",4:"deer",5:"dog",6:"frog",7:"horse",8:"ship",9:"truck"}
for i in range(len(date_dict)):
    path1 = f"train/{date_dict[i]}"
    path2 = f"test/{date_dict[i]}"
    if not os.path.isdir(path1):
        os.makedirs(path1)
    if not os.path.isdir(path2):
        os.makedirs(path2)    

#生成训练集图片，如果需要jpg格式，只需要改图片后缀名即可。
for j in range(1, 6):
    dataName = "data_batch_" + str(j)  # 读取当前目录下的data_batch12345文件，dataName其实也是data_batch文件的路径，本文和脚本文件在同一目录下。
    Xtr = unpickle(dataName)
    print(dataName + " is loading...")

    for i in range(0, 10000):
        img = np.reshape(Xtr[b'data'][i], (3, 32, 32))  # Xtr['data']为图片二进制数据
        img = img.transpose(1, 2, 0)  # 读取image
        picName = 'train/' + str(date_dict[Xtr[b'labels'][i]])+"/"+str(i + (j - 1)*10000) + '.png'  # Xtr['labels']为图片的标签，值范围0-9，本文中，train文件夹需要存在，并与脚本文件在同一目录下。
        imsave(picName, img)
    print (dataName + " loaded.")

print("test_batch is loading...")

# 生成测试集图片
testXtr = unpickle("test_batch")
for i in range(0, 10000):
    img = np.reshape(testXtr[b'data'][i], (3, 32, 32))
    img = img.transpose(1, 2, 0)
    picName = 'test/' + str(date_dict[testXtr[b'labels'][i]]) + '/' + str(i) + '.png'
    imsave(picName, img)
print("test_batch loaded.")

参考链接：https://blog.csdn.net/guohuifengby/article/details/62424299
参考链接：https://blog.csdn.net/weixin_44633882/article/details/86905285