Python计算两图相似性-SSIM、PSNR，MSE

1、简介

SSIM：值越接近1，图像越相似

PSNR：PSNR越大说明失真越少，生成图像的质量越好

MSE：MSE值越小，图像越相似

 

2、代码示例

测试图片点击进行下载：Image

# -*- coding:UTF-8 -*-
from skimage.metrics import structural_similarity as SSIM
from skimage.metrics import peak_signal_noise_ratio as PSNR
from skimage.metrics import mean_squared_error as MSE
import cv2

def get_spm(img_cp1, img_cp2):

    psnr = PSNR(img_cp1, img_cp2)
    ssim = SSIM(img_cp1, img_cp2, multichannel=True, channel_axis=2)
    mse = MSE(img_cp1, img_cp2)
    return psnr, ssim, mse

img_cp1 = cv2.imread("WD1.png")
img_cp2 = cv2.imread("WD2.png")
psnr, ssim, mse = get_spm(img_cp1, img_cp2)
print("PSNR:{}\nSSIM:{}\nMSE:{}".format(psnr, ssim, mse))