Idea (not unique): Using deconvolved images for viewing

The idea was to use deconvoluted images to make an image sharper when viewed by a near or far sighted person. Could make distant signs sharper or help diagnose eye problems. For instance, a person with eye problems could sit in front of a screen (at a known distance, far or close). Deconvolved images could be displayed on the screen where the person would press buttons indicating which deconvolution resulted in the sharpest image. This would be done iteratively to eventually estimate the point spread function of their eye only using a computer screen (might be useful for countries that couldn’t afford regular vision diagnostic tools).

First part has been done: http://dsplab.eng.fiu.edu/DSP/Research/research_deconvolution_project.html
Pretty cool stuff. With the example image, you can blur your eyes or use an out of focus camera to convolve and restore the shape of the letters. I think the lack of clarity comes from the fact that they used full intensity black and white images…which means there will absolutely be artifacts since the screen can’t have enough dynamic range.

I’m sure using it for diagnostics has been done…just don’t know optometry terms.