Calculating the RMSE value between two images

2022 9 19
1 个回答
回答已采纳
更新时间：2022 9 20
26 次查看（30 天）

0 个投票

I calculated the RMSE value between two images of size 128 x 128 in .jpg format and I am getting the result above 1. I want the range to be between 0-1. I tried applying imadjust to normalize but it did not help. How can I normalize the image and get the required result between 0-1. I also tried to save the images in .mat file and normalize it but I am getting error as the image is being saved as type struct and not double. Please help.Thank you in advance.
I used the following line to calculate rmse:
rmse_val = sqrt(mean((ref(:)-a(:)).^2))

请先登录，再回答此问题。

 采纳的回答

Image Analyst
Image Analyst 2022-9-19
Ran in:

1 个投票

Ran in:
You forgot to attach your variables. I don't know if you can subtract structures like that so I think ref and (the badly-named) a are actually numerical arrays, not structures. If they are uint8 gray scale images, you should convert to double before subtracting, and use the rms function:
a = imread('cameraman.tif');
ref = a + 10;
% First method is to use rms on the difference array.
diffImage = double(ref) - double(a);
r = rms(diffImage(:))
r = 9.9975
% Alternate way is to use rmse
re = rmse(double(ref(:)), double(a(:)))
re = 9.9975

4 个评论
显示 2更早的评论 隐藏 2更早的评论

Shourya
Shourya 2022-9-20
@Image Analyst Thank you for the solution and explaination.Thank you for pointing out the variables details. Sorry I missed them. The groundtruth image is referred to as 'ref' and the test image is referred to as 'a'. Yes, they are uint8 gray scale images. I tried the above solution but I get the result as 5.69. I want it between 0-1 or how can I get the result in precentage.
Image Analyst
Image Analyst 2022-9-20
You can divide either by the max value in your reference image, or divide by 255, depending on how you define percentage.
Shourya
Shourya 2022-9-20
This is what I will get 5.69/255=0.0219. Is this right?

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Image Processing Toolbox 的更多信息

产品

标签

Community Treasure Hunt

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

Start Hunting!

Translated by