Once upon a time (years ago) i wrote a tool called bookbuilder, which did exactly this :-) Take a bunch of camera photos of a book on dark background, find the

Once upon a time (years ago) i wrote a tool called bookbuilder, which did exactly this :-) Take a bunch of camera photos of a book on dark background, find the edges and extract the text part :-) I'm not sure, it is still working, because it was java 6... but it is using the excellent boofCV library (http://boofcv.org http://boofcv.org) If you would like to try it, give it a shot (tesseract 3 is needed): https://mega.nz/#!EVxA0ZoD!6Uy5A4HexJbewXqkOLsW-4sj5IO5LOGefTfN0z-acQE https://mega.nz/#!EVxA0ZoD!6Uy5A4HexJbewXqkOLsW-4sj5IO5LOGef... java -jar bookbuilder-0.2.jar --input-path=input --output-file=output/output.pdf --ocr-embed-layer --rotation-degrees=180