Browse Teach Curate

Automated Manhwa Translation Pipeline for Low-Resource Languages

How comic translation can be automated end-to-end by chaining an object detector that finds speech bubbles, OCR that extracts the text, a neural machine translation model fine-tuned for a low-resource language pair, and re-overlay of the translated text onto the artwork. Illustrates the design decisions and evaluation metrics involved when composing vision and language models into one multimodal workflow.

Chief Galen Tyrol Text

Crossing Language Borders: A Pipeline for Indonesian Manhwa Translation Nithyasri Narasimhan and

In this project, we develop a practical and efficient solution for automating the Manhwa translation from Indonesian to English. Our approach combines computer vision, text recognition, and natural l…