2
$\begingroup$

I am trying to increse the quality of the images that I gather from the microscope. That is a acoustic microscope and there are lots of technical details but in a nutshell the low quality images and its corresponding high quality images that I gather from the same sample are not perfectly alligned because in my setting it is impossible to increase quality without removing the sample from the microscope so when I put it again it is a manual process so they are not perfectly aligned.

Output of my network will be,let's say, 256 x 256 image and its corresponding label will be high quality 256 x 256 image,in theory, of the exactly same area. If I make pixel to pixel comparison between them, for example taking MSE for the loss function, will it be able to learn ? I am not sure because pixels are not perfectly alligned, they do not represent the same area of the image(the difference is not that great but they are not perfectly alligned as I said)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.