I have written three codes but I am not sure which one is correct?
My goal is to measure the MSE (Mean Squared Error) between two images (reduced color and original image)
I want to measure MSE of RGB components and Lab components between image 1 and 2? As the recent publications show the MSE value in unsigned integer 8 bit, I want to apply the correct command for unsigned integer 8 bit [0, 255].
I am farshid keivanian, MSc in final semester, near to defend process of thesis.
Electronics Engineering, Birjand,
Thanks
See, e.g., http://stackoverflow.com/questions/13956051/matlab-how-to-obtain-values-of-pixels
Why did You wrote such code?
if R0_255(j,i) > R2_0_255(j,i)
d1=(R0_255(j,i)-R2_0_255(j,i))^2;
elseif R2_0_255(j,i) > R0_255(j,i)
d1=(R2_0_255(j,i)-R0_255(j,i))^2;
elseif R2_0_255(j,i) == R0_255(j,i)
d1 = 0;
end
May be this one code line is sufficent?
d1=(R0_255(j,i)-R2_0_255(j,i))^2;
Some approaches suggest to transform the image in YCbCr space, once you did this transformation you could use Luma (Y) channel for MSE measure in a full reference schema (For example, the original image and the noisy version).
You may refer a more better solution to this:
Chapter A Novel Full-Reference Image Quality Index for Color Images
Dear Farshid Keivanian,
Look the link. May be useful.
http://research.ijcaonline.org/volume29/number9/pxc3874978.pdf
Regards, Shafagat
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification