not working with large amount of data

[akshat9425]

hi,

i need to implement ocr for some financial docs this repo works well but only for few lines not for a complete table please tell me how could i change repo for below type documents

thanks

[hussaintamboli]

Hi. I worked on this repo a long time ago. But as I remember, to recognise the text more accurately you need to train the tesseract (instructions here).

By looking at your document, I think you also need to parse the generated text afterwards to get the values.