MyWSL: Malaysian words sign language dataset

Area of Science:

• Computer Science
• Linguistics
• Human-Computer Interaction

Background:

• Sign language is the primary communication method for deaf and hard-of-hearing individuals.
• A lack of accessible datasets hinders research in Malaysian Sign Language (MySL) recognition and translation.
• Communication barriers exist for non-signers interacting with MySL users, particularly within families.

Purpose of the Study:

• To introduce and make publicly available the MyWSL2023 dataset, the first image dataset for Malaysian Sign Language.
• To support the development of automated systems for MySL recognition and translation.
• To facilitate research in vision-based gesture and sign language recognition.

Main Methods:

• Collected 3,500 static images of ten common Malaysian Sign Language words.
• Data captured from five participants (2 male, 3 female, aged 20-21) indoors under normal lighting.
• Dataset named MyWSL2023, made freely accessible for research.

Main Results:

• The MyWSL2023 dataset contains 3,500 images of ten static MySL words.
• The dataset was collected under controlled indoor lighting conditions.
• The dataset is now publicly available for the research community.

Conclusions:

• The MyWSL2023 dataset is a valuable resource for advancing Malaysian Sign Language research.
• This dataset will aid in the development of automated sign language recognition systems.
• Facilitates further investigation into vision-based gesture recognition for MySL.