For internal purposes we print pre filled PDF documents which are then completed manually (hand-writing) and scanned again. We use this documents for validating internal education programs. The education program as pre-printed in the PDF.
We experience a lot of problems recognizing the documents because of bad scanning or bad recognition (see attached document). What would be the best strategy to classify these documents? We always use a cover sheet (fully customizable) as well as some attachments (details about the course) which can very in stricture but are never recognized.
Till now we recognize the word "Formular" written on the left upper corner as well as the grid lines used to fill in text. Unfortunately this does not work well enough. My questions are:
- What are your recommendations?
- Is a QR-Code in the upper corner a better idea?
- How to detect scaling problems (see attached image)?
- Document definition recognition is done using the word "Formular" and some grid lines. Any improvements? Would you add special corners?
- Any general advice?