I think you have to distinguish between noise and distortion. The distorted or blurred image requires a special restoration technique like " deconvreg". But for a noisy image, SNR could be calculated, and different types of filters may be used to denoise it. The question; for which reason you calculate the SNR for your distorted binary image! In this case, it is normal to get high SNR.

Try this. You may get

InputImage=imread(‘Input.jpg’); ReconstructedImage=imread(‘recon.jpg’); [M N]=size(InputImage); MSE = sum(sum((InputImage-ReconstructedImage).^2))/(M*N); PSNR = 10*log10(256*256/MSE); % 256 for 8- bit image

PSNR = 10*log10(2*2/MSE); % 2 for binary image

The PSRN is mathematical function. It is used as quantitative evaluation. In image processing algorithm validation, both subjective and objective evaluation are necessary. For your case, the PSRN is improved but quality is deteriorate so it is not good for visual application.

Hi, as I understand you are comparing the original inserted binary mark and the extracted one for a watermarking scheme? If it is the case i advise you to use the correlation and the BER(Bit error rate) metrics to measure the difference between them. In fact, they give accurate results than the PSNR.

Sama Lenin Kumar Reddy i used the second equation for binary image the psnr is less than 8 dB...

Note that, generally, PSNR is relative and based answers above as well as your responses it would be better to consider other image quality metrics, notably the bit error rate (BER).

All the best!

PSNR is used to measure the noise ratio between original and estimated image

Try to use metrics like SSIM, NCC, BER etc. They are more accurate.