OCRing YouTube Music with Common Lisp: A Pixel-Perfect Adventure
2025-01-06
A developer attempted to extract music data from a YouTube video using Common Lisp. Initial attempts with Tesseract and ChatGPT proved unsuccessful. Ultimately, an old-school pixel differencing method, involving manual extraction of character images and comparison, successfully extracted most of the musical notation. While not perfect, the extracted data sufficed, proving the method's feasibility. The article also details the developer's experience using Lisp for image processing and efficient development.
Read more
Development
Image Processing