Form recognition software
I am looking for form recognition SW for Linux, preferably open source. I have about 500,000 picture files of scanned questionnaires and I need to extract data from these. The data are either bar codes (most important), tick marks or handwritten numbers and text. The questionnaires has alignment marks that can be used for orientation. Does anyone know of any software that can perform this task? I have tried programs like tesseract, but it will read all text on the page. OCRopus seem promising but it isn't there yet.