Related Experiment Video
Updated: Jul 8, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.5K
KU-BdSL: An open dataset for Bengali sign language recognition
Abdullah Al Jaid Jim1, Ibrahim Rafi1, Md Zahid Akon2
1Electronics and Communication Engineering Discipline, Khulna University, Khulna 9208, Bangladesh.
Data in Brief
|December 11, 2023
Summary
Researchers developed the KU-BdSL dataset for Bengali sign language (BdSL) recognition. This computer vision dataset aids communication for the deaf and dumb community, enabling Bengali consonant identification.
Area of Science:
- Computer Science
- Artificial Intelligence
- Linguistics
Background:
- Sign languages, like Bengali sign language (BdSL), present communication challenges due to complex patterns.
- Machine learning and computer vision offer potential solutions for bridging communication gaps for speech and hearing-disabled individuals.
Purpose of the Study:
- To create and introduce the KU-BdSL dataset for recognizing Bengali consonants within BdSL.
- To facilitate the development of machine translation and human-machine interfaces for the deaf and dumb community.
Main Methods:
- Collected 1500 images of hand signs representing 38 Bengali consonants ('banjonborno').
- Dataset named `KU-BdSL' comprises 30 classes.
- Images were captured using smartphones by 39 participants.
Main Results:
- The KU-BdSL dataset contains 1500 images across 30 classes, focusing on Bengali consonants.
- The dataset enables the identification of Bengali consonants from images and videos.
Conclusions:
- The KU-BdSL dataset is a valuable resource for advancing Bengali sign language recognition.
- This dataset can support the development of assistive technologies for the deaf and dumb community and future work on BdSL vowels and words.
