It’s more a matter of algorithms - I’m not very mathematical, so I was looking for an engineering solution ... If this is disconnected from the topic for SO, let me know and I will delete the question.
I created an open source mashup to do optical character recognition on complex backgrounds: https://github.com/metalaureate/tesseract-docker-ocr
I want to use it to scan shortcuts with a predefined identification code, for example, 2826672. The accuracy is about 70% for numbers.
Question: how to add redundancy code in my code to increase accuracy to 99%, and how to decode it? I can imagine some really trivial ways, such as doubling and inverting numbers, but I don’t know how to do this in such a way as to distinguish information theory without having to translate a lot of mathematics.
How to add and decode numbers to correct OCR errors?
source
share