As part of one of the business tasks, it is necessary to recognize scans of printed and signed documents. Which OCR technology is preferred for this task and provides a higher percentage of correct recognition? Or maybe some kind of pre-processing of scans such as highlighting or increasing contrast may be necessary? But I would also like to do this with a robot inside the process
Hi, @markuss! We can’t recommend anything specific - we need to look at specific examples. Also, the accuracy of recognition depends on many factors, including the quality of scanning and the ability of developers to process the information obtained. No pre-processing of scans is required for the recognition of standard documents.
Printing is a complicating factor, but it is necessary to check how much. In this case it is impossible to tell for sure about realization, it is required to be creative. You can use a third-party service, for example, to replace the blue color of the print in the document with white. This operation can be automated and performed within the framework of the robot algorithm.
@markuss, did everything become clear to you after the answer? Please ask questions if your situation is not completely resolved.
If everything is clear, please, mark my previous post as the “Solution” to the question in the topic.