My main goal is to create a set of tools which help in cleaning images to make them suitable for printing and viewing. Here's a high level view of some of the operations needed: Grayscale Scan -> deskew -> bilevel -> clean -> output image Everything is pretty experimental right now. Supported image types: ---------------------- For now only pbm and pgm files are supported. In the future I'd like to support tiff reading and writing. This would also help toward a goal of creating a full pdf as output. Deskew: ------- The Hough Transform is used to find the skew angle. It isn't particularly fast, but for now it is quick enough and it works if the pages isn't very skewed. Bilevel: -------- There are many methods for converting grayscale images to bilevel images. I'm still experimenting to find some good defaults. I want to use and adaptive method to cut out as much noise and inner page curl as possible. Clean: ------ The current method segments the image into contiguous sections, blobs, of either white or black pixels. Blobs which contain fewer pixels than some given threshold are discarded. More cleaning method need to be researched.