如何计算自己的图片数据集的均值和方差

最新推荐文章于 2025-03-16 20:33:10 发布

原创最新推荐文章于 2025-03-16 20:33:10 发布 · 7k 阅读

18 ·

本内容遵循CC 4.0 BY-SA版权协议

收录于

常用代码

本文介绍了一个Python脚本，用于从CIFAR-10数据集中加载图片并计算RGB通道的平均值和方差。通过遍历数据集中的每个图像，该脚本计算每个图像的RGB通道平均值和标准差，然后求出整个数据集的RGB平均值和方差。

import numpy as np
import pickle
def load_CIFAR_batch(filename):
    """ load single batch of cifar """
    with open(filename, 'rb') as f:
        datadict = pickle.load(f,encoding='latin1')
        X = datadict['data']
        Y = datadict['labels']
        X = X.reshape(10000, 3, 32,32).transpose(0,2,3,1).astype("float")
        Y = np.array(Y)
        return X, Y
X,Y = load_CIFAR_batch('/home/xm/桌面/thesis_temp/AdaBound-master/demos/cifar10/data/cifar-10-batches-py/test_batch')
img = X
R_means = []
G_means = []
B_means = []
R_stds = []
G_stds = []
B_stds = []
for im in img:
    im_R = im[:,:,0]/255
    im_G = im[:,:,1]/255
    im_B = im[:,:,2]/255
    im_R_mean = np.mean(im_R)
    im_G_mean = np.mean(im_G)
    im_B_mean = np.mean(im_B)
    im_R_std = np.std(im_R)
    im_G_std = np.std(im_G)
    im_B_std = np.std(im_B)
    R_means.append(im_R_mean)
    G_means.append(im_G_mean)
    B_means.append(im_B_mean)
    R_stds.append(im_R_std)
    G_stds.append(im_G_std)
    B_stds.append(im_B_std)
a = [R_means,G_means,B_means]
b = [R_stds,G_stds,B_stds]
mean = [0,0,0]
std = [0,0,0]
mean[0] = np.mean(a[0])
mean[1] = np.mean(a[1])
mean[2] = np.mean(a[2])
std[0] = np.mean(b[0])
std[1] = np.mean(b[1])
std[2] = np.mean(b[2])
print('数据集的RGB平均值为\n[{},{},{}]'.format(mean[0],mean[1],mean[2]))
print('数据集的RGB方差为\n[{},{},{}]'.format(std[0],std[1],std[2]))

标签

#图片数据集预处理