How can I calculate the MSE between 2 images?

24 次查看（过去 30 天）
Tanish Mehta
Tanish Mehta 2019-10-31
编辑： KALYAN ACHARJYA 2019-11-5
ref=imread("birdmiddle.png");
A=imread("birdmiddle2.png");
err = immse(A, ref);
fprintf('\n The mean-squared error is %0.4f\n', err);
** I am trying to find the MSE but the following error shows. Image 'A' is the original image 'ref' compressed using lossy jpeg compression, and then the compressed jpeg is converted back into a png. Why is there an error?**
Error using immse (line 34)
A and B must have the same class.

请先登录，再回答此问题。

回答（1 个）

KALYAN ACHARJYA
KALYAN ACHARJYA 2019-10-31
编辑：KALYAN ACHARJYA 2019-10-31
ref=imread('birdmiddle.png');
A=imread('birdmiddle2.png');
err = immse(A, ref);
fprintf('\n The mean-squared error is %0.4f\n', err);
Still problem is not solved, please attached two images
  2 个评论
Tanish Mehta
Tanish Mehta 2019-11-1
I think the problem earlier was that the original picture was uint16 but the second image was uint8. So i converted the second image to uint16 so both images would have the same class. The MSE calculation was successful however the number is very high. The MSE was '2041744502.9629'. Can this be correct?
Thanks a lot for the help!
KALYAN ACHARJYA
KALYAN ACHARJYA 2019-11-5
编辑：KALYAN ACHARJYA 2019-11-5
Can you attach the two original images? One suggestion, convert the both to uint8 and find the same, do share the result value (MSE)?
psnr-and-mse.png
Tanish do calculate (samples) on notebook, you get the idea.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Image Quality 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by