YouTube Hardcoded Subtitle OCR Userscript

Overview

This userscript is designed to help you extract hardcoded subtitles from YouTube videos using Optical Character Recognition (OCR). The script provides an option to crop a specific area of the video for OCR processing.

Features

• Crop Area Selection: Select a specific area of the video to focus the OCR process.
• Automatic Pausing: The video automatically pauses during the OCR process to ensure accurate text recognition.

Requirements

Userscript Manager for browser, visit greasyfork for installation instructions. [made with tampermonkey]

Installation

1. Open browser and enable extension

2. Download and install YouTube Sub OCR JP UserScript

   • Download from GitHub jp user.js

Demo

Usage

• Open a YouTube video.

• Click on the Crop button below video canvas to select the area for OCR.

• Click the OCR button to begin the OCR process.

• The video will pause, and the selected area will be processed for text extraction.

• The recognized text will be displayed below the video canvas.

• Use browser extension such as 10Ten or Yomitan to read the characters and make flashcards for studying with.

Technical Details

1. The OCR is done by tesseract
2. The crop area is done by cropperjs