Simple OCR Hardsubs

--> Extract hardsub to images using InpaintDelogo
--> detect text area using Cloud Vision API and clean images
--> Ocr using Cloud Vision API

Setup

• Install Python
• run cmd pip install -r requirements.txt
• Install AviSynth+ : Link
• Recommend using AvsPmod to edit avs script
• ffmpeg for window link
• Install InpaintDelogo plugin: Link
• Everything else 😎 Linkhttps://cloud.google.com/vision/docs/detect-labels-image-client-libraries

Usages

Drop videos to create_avs_script.py -> create avs script

Drop folder to gg_clean_image_text.py -> clean images

Drop folder to ocr gg seperate image.py -> OCR

Run create srt from txt files.py -> Make srt files (it's remove some japanese lines)

You can change parameters by editing python script.

😸