Scanning videos so that I can see what text is shown. The Tesseract OCR is kind of okay, didn’t get the best results but my text on 1080p video is quite small so it might not be the program’s fault.
Made a script with ChatGPT that takes a folder of images, analyzes them with Tesseract, then makes multiple text files of the extracted text.
Then I updated it to make sure that it makes one file with no duplicate text. That way, it’s not recording text from multiple frames.
Here’s the final repo