Matlab codes to find MSE of many images against one

I know how to get the MSE(Mean Square Error) of one image against one image, but how can i do it to get the MSE of many images against one image which serve as the template?

The Matlab code below is the i use to get the MSE of one image against one image, how can i adjust it to solve my problem?

InputImage=imread('7-6.jpg');
ReconstructedImage=imread('7-6a.jpg');
n=size(InputImage);
M=n(1);
N=n(2);
MSE = sum(sum((InputImage-ReconstructedImage).^2))/(M*N)
PSNR = 10*log10(256*256/MSE)
fprintf('\nMSE: %7.2f ', MSE);
fprintf('\nPSNR: %9.2f dB', PSNR);

Answer 1

[KALYAN ACHARJYA]

MSE_data=[];
PSNR_data=[];;
Images=dir('C:\Complete_path\Input_folder_name\*.png');
%................Input Images, Note on .format...^^
ReconstructedImage=imread('7-6a.jpg');
for i=1:length(Images)
ImgName=strcat('C:\Complete_path\Input_folder_name\',Images(i).name);
InputImage=imread(ImgName);
n=size(InputImage);
M=n(1);
N=n(2);
MSE_data(i)= sum(sum((InputImage-ReconstructedImage).^2))/(M*N);
PSNR_data(i)= 10*log10(256*256/MSE(i));
fprintf('\nMSE: %7.2f ', MSE_data(i));
fprintf('\nPSNR: %9.2f dB', PSNR_data(i));
end

Also, here you are saving MSE and PSNR of all images in arrays, you can acess it later in your code.

[KALYAN ACHARJYA]

Hope the question is answered!

[Abdulmu'min Musa]

Not completely, i am still having some errors. From the code you send to me, this is how i adapt it to mine.

MSE=[];
PSNR=[];
Images=dir('C:\Users\ENGR. KAYODE\Desktop\T6MSE\*.jpg');
InputImage=imread('7-1.jpg');
for i=1:length(Images)
ImgName=strcat('C:\Users\ENGR. KAYODE\Desktop\T6MSE\',Images(i).name);
ReconstructedImage=imread(ImgName);
n=size(InputImage);
M=n(1);
N=n(2);
MSE(i)= sum(sum((InputImage-ReconstructedImage).^2))/(M*N);
PSNR(i)= 10*log10(256*256/MSE(i));
fprintf('\nMSE: %7.2f ', MSE(i));
fprintf('\nPSNR: %9.2f dB', PSNR(i));
end

When i run it, i was encountering errors, and when i copy each step and paste it on command window to run, this is what i got.

At (ReconstructedImage=imread(ImgName))

the error was

Error using fopen

Filename is too long.

Error in imread (line 343)

fid = fopen(filename, 'r');

and at (MSE(i)= sum(sum((InputImage-ReconstructedImage).^2))/(M*N))

the error was

Subscripted assignment dimension mismatch.

and at (PSNR(i)= 10*log10(256*256/MSE(i)))

the error was

Index exceeds matrix dimensions.

Pls, help me out.

[KALYAN ACHARJYA]

InputImage and ReconstructedImage are same sizes, yes?

[Abdulmu'min Musa]

Yes.

185 * 150

[KALYAN ACHARJYA]

Please change the MSE and PSNR variable name to different, as I provided in my answer, so the conflict with inbuilt function can be avoided.

Also pleased do attach one sample image and 7-1.jpg, so that I can try on it

[Abdulmu'min Musa]

Sorry for the late reply, had problem with the network. So, there they are.

Those without (a) are the original image e.g 1-1.jpg, 1-2.jpg and those with (a) are the enhanced image e.g 1-1a.jpg, 1-2a.jpg, etc.

[KALYAN ACHARJYA]

Always read the comment, what I have asked for "do attach one sample image" one sample image and ReconstructedImage, why you are attaching so many?

Which one is the ReconstructedImage?

[Abdulmu'min Musa]

Sorry pls, i thought of attaching many so as to give more rooms for options.

The recontructed image are 1-1a, 1-2a, 1-3a, 1-4a, 1-5a. Thanks

[KALYAN ACHARJYA]

When I tried with single image its working, the MSE results is zero shows because both Input Image and reconstruucted images are same

ReconstructedImage=rgb2gray(imread('1-1a.jpg'));
InputImage=rgb2gray(imread('1-1a.jpg'));
[M N]=size(InputImage);
error=InputImage-ReconstructedImage;
MSE_data=sum(sum(error.* error))/(M * N);
PSNR_data=10*log10(256*256/MSE_data);
fprintf('\nMSE: %7.2f ', MSE_data);
fprintf('\nPSNR: %9.2f dB', PSNR_data);

Considering 1-1.jpg image as input image and i-1a.jps as reconstructed image. I dont think, there is any images calling loop issue, requested you to verify the path correctly

Command Window:

MSE:    0.00  % The reason of zero MSE already explained
PSNR:       Inf dB>>

Also you can use inbuilt MSE and PSNR function also.

[KALYAN ACHARJYA]

Here updated

MSE_data=[];
PSNR_data=[];;
Images=dir('C:\Complete_path\Input_folder_name\*.png');
%................Input Images, Note on .format...^^
ReconstructedImage=rgb2gray(imread('7-6a.jpg'));
for i=1:length(Images)
ImgName=strcat('C:\Complete_path\Input_folder_name\',Images(i).name);
InputImage=rgb2gray(imread(ImgName));
[M N]=size(InputImage);
error=InputImage-ReconstructedImage;
MSE_data(i)=sum(sum(error.* error))/(M * N);
PSNR_data(i)=10*log10(256*256/MSE_data(i));
fprintf('\nMSE: %7.2f ', MSE_data(i));
fprintf('\nPSNR: %9.2f dB', PSNR_data(i));
end

[Abdulmu'min Musa]

Thank you so so so much. It works.

I appreciate all your patience and efforts in helping me out.

Once again, thanks.

[KALYAN ACHARJYA]

Always Welcome @Musa

[simarjit kaur]

this code correctly run bt not show any output

plz hlp me sir

[KALYAN ACHARJYA]

Please do share (detail).