kprateek7
commented
3 years ago Hi! Assuming you really are using it in global shutter mode (some software have it set to rolling shutter by default), these images make a case for shortening the exposure time. The typical times for our work are below 0.2 millisecond and the translation rate is limited to a few millimeter per second. Your images look like they have been exposed for too long (or moving too fast) and I suggest you work on that, because no amount of deblur will give you good images this way (unless you use some fancy deep learning). Let us know if you get it to work better Prateek
SergeTauger
Thank you for your work! Do I get correctly that the code is used only to remove residual focus error? Could you please recommend a way to remove motion blur in continous whole slide imaging? I cannot use simple deconvolution with line kernel since the kernel seems not be like line, and I don't know why. The images (still image, motion-blurred image taken at constant speend with global shutter CMOS camera, and PSF obtained from the first two using naive Wiener filtration) can be found here.