This code not correct result the what is problem ?
Mostrar comentarios más antiguos
MSE1=mean(mean((embeddedImage-originalImage).^4));
MaxI=1;% the maximum possible pixel value of the images.
PSNR=15*log10((MaxI^2)/MSE1);
disp(PSNR);
1 comentario
the cyclist
el 27 de Abr. de 2015
Please read this, and then edit your question to have a better chance at getting a useful answer: http://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer
Respuestas (1)
Image Analyst
el 27 de Abr. de 2015
0 votos
Chances are your images are uint8 and the subtraction is clipping them. Cast to double before subtracting. And I'm not sure why you're raising them to the 4th power instead of squaring them to find the Mean Square Error. Finally, you can use mean2() instead of mean(mean)).
Better yet, just use the built-in psnr() function in the Image Processing Toolbox. Or, you can use my attached function if you have an antique version of MATLAB.
Categorías
Más información sobre Image Quality en Centro de ayuda y File Exchange.
Productos
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!