Sin-Cap: Automatic Sinhala Audio Captioner using Neural Networks

Abstract

Audio captions or video subtitles have been extremely beneficial not only for deaf or hard of hearing peopl, but also for related language illiterate people. And when in situations as hard to focus and hard to listen, captions are greatly beneficial. But even if there are so many automated captioning systems for high resource languages, there are less accurate and reliable systems for low resource languages. An end-to-end automated Sinhala captioning system is the proposed solution when considering the existing systems and requirements because even if captioning systems are available for Sinhala, there are many drawbacks to them. This research is carried out to develop a better solution for the low- resource language.