批量计算PSNR和SSIM

最新推荐文章于 2024-03-20 15:41:26 发布

原创最新推荐文章于 2024-03-20 15:41:26 发布 · 2.6k 阅读

28 ·

CC 4.0 BY-SA版权

文章标签：

#python #tensorflow

code tips 同时被 2 个专栏收录

30 篇文章

订阅专栏

11 篇文章

订阅专栏

本文介绍如何使用TensorFlow计算图像间的峰值信噪比（PSNR）和结构相似性指数（SSIM），通过读取测试图片并运行在指定文件夹中的图像对，计算并记录每张图片的平均PSNR和SSIM，展示了在81-100张图片中的性能指标和总耗时。

部署运行你感兴趣的模型镜像

import tensorflow as tf
import numpy as np
import time

start = time.clock()

def read_img(path):
    return tf.image.decode_image(tf.read_file(path))

def psnr(tf_img1, tf_img2):
    return tf.image.psnr(tf_img1, tf_img2, max_val=255)

def ssim(tf_img1, tf_img2):
    return tf.image.ssim(tf_img1, tf_img2, max_val=255)

def main(a, b):
    # t1 = read_img('./Data/00001_00_0.1s_gt.png')               #Clean Image
    # t2 = read_img('./Data/00001_00_0.1s_out_optimized.png')       #Noise Image
    t1 = read_img(a)
    t2 = read_img(b)
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        p = sess.run(psnr(t1, t2))
        s = sess.run(ssim(t1, t2))
    return p, s

path1 = './images/test1/'  # 指定输出结果文件夹
path2 = './images/test2/'  # 指定原图文件夹
list_psnr = []
list_ssim = []
file = open(r'计算结果.txt', mode='w', encoding='utf-8')
for i in range(81, 101):
    print('第%s张图片' % i)
    if i == 100:
        a = path1 + str(i) + '.png'
        b = path2 + str(i) + '.png'
    else:
        a = path1 + '0' + str(i) + '.png'
        b = path2 + '0' + str(i) + '.png'
    p , s = main(a, b)
    list_ssim.append(s)
    list_psnr.append(p)
    file.write('第%s张图片:' % i + '\n')
    file.write('psnr, {:.5f}'.format(p) + '\n')
    file.write('ssim, {:.5f}'.format(s) + '\n')


elapsed = (time.clock() - start)
file.write("\n")
file.write("汇总:\n")
file.write('平均PSNR, {:.5f}'.format(np.mean(list_psnr)) + '\n')
file.write('平均SSIM, {:.5f}'.format(np.mean(list_ssim)) + '\n')
file.write('Time used, {:.5f}'.format(elapsed) + '\n')
file.close()