Hacker Newsnew | comments | show | ask | jobs | submitlogin

It's quite possible. Not only is it possible to perform deconvolution, but since you know that you're looking for text data, and you even have the font, you can do much better. You can iterate through millions of names very quickly and find which one, once blurred produces the best match.

In the same vein, check out http://www.ee.columbia.edu/~wliu/CVPR05_LiuWei1.pdf

Applications are open for YC Summer 2015

Guidelines | FAQ | Support | Lists | Bookmarklet | DMCA | Y Combinator | Apply | Contact