This repository contains the codes and resources for the project "Automatic Subtitle Generation for Bengali Multimedia using Deep Learning." This project aims to enhance the accessibility and inclusivity of Bengali multimedia by automatically generating subtitles using state-of-the-art deep learning techniques.
- 🎧 Automatic Transcription: Convert Bengali audio to text using the Wav2vec2 model.
- 📚 Common Voice Bengali Dataset: Leverage a large collection of Bengali audio recordings.
- ⏰ Synchronization: Sync transcribed text with corresponding audio segments.
- 📝 Post-Processing: Enhance readability with capitalization and punctuation restoration.
- 🌐 Improved Accessibility: Enhance the viewing experience for Bengali multimedia audiences.
-
Authors:
- Ehsanur Rahman Rhythm (@errhythm) - Conceptualization, Coding, Visualization, Writing
- Shafakat Sowroar Arno - Writing (Editing)
- Rajvir Ahmed Shuvo - Writing (Editing)
-
Supervisors:
- Sifat E Jahan
- Annajiat Alim Rasel
I would like to extend my heartfelt thanks to everyone who encouraged and motivated me throughout this project. A special thank you to the insightful YouTube comment that sparked the initial idea for this project. I am also deeply grateful to my thesis teammates for their collaboration and for allowing me to focus on the parts I contributed which was basically the whole project.