I am looking to build a for fun project for my work during the holidays (I work at a recreational basketball centre), that takes in scanned images of social basketball competition scoresheets and scans them using an OCR. With this scan, I want to take in the scores and fouls committed by players and put them into a database.

In the future, I want to analyse this database and make a website/app where I can display the competition statistics such as Leading Points scorer, most fouls committed and other interesting basketball statistics.

I have only basic programming experience from currently doing a physics degree (undergrad) and looking to learn some programming (don’t mind any language) and wanna do it through a fun project.

My main difficulty with this project has been OCR libraries having difficulties recognising hand written scoresheets formats (in particular player names and strikes to denote a score). I will add an example scoresheet format at the end here and a link to pdf where you can learn how FIBA scoresheets work! Link to PDF: https://collingwoodbasketball.com.au/wp-content/uploads/2020/08/FIBA_scoresheet_quarters.pdf

enter image description here

So far I have tried python libraries such as Tesseract and EasyOCR but they both have quite a few difficulties recognising the handwritten solutions. Maybe I am going in the wrong direction because ChatGPT’s way of using its OCR can actually recognise game scores and player names etc., and get useful information from the scanned images.

Instead of recognising the positions of text (x, y and height, which is what I ended up doing lol since im a noob), I want to learn how to put player names, points and fouls into there own dataframes with the database.

Could someone please point me in the right direction.