Citation

Min, Pa Pa and Kua, Janielle Ai Lyinn (2026) Deep Learning-Based Recognition of Malaysian Sign Language Gestures. In: 14th International Conference on Information and Communication Technology, ICoICT 2026, 30 July 2026 - 31 July 2026, Bandung.

Text 13.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

The goal of this research is to enhance communication for the deaf and speech-impaired community by developing a deep learning-based Malaysian Sign Language recognition system. The suggested model uses a hybrid Convolutional Neural Network and Long Short-Term Memory architecture to classify a range of static MSL gestures, including alphabets, numbers, and simple Malay words. It does this by using two inputs: hand keypoint heatmaps and grayscale images. Datasets from various sources were gathered, and preprocessed using techniques such as data augmentation, background removal, grayscale conversion, and keypoint extraction with MediaPipe to ensure reliable performance. Two separate datasets—Dataset 1 and Dataset 2,—were used to train and evaluate the system. Experimental results show that the model achieves 97% classification accuracy before fine-tuning and 99% after, with high precision, recall, and F1-scores across gesture classes. The model’s performance surpasses that of existing methods in both consistency and accuracy across gesture classes.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Malaysian Sign Language , gesture recognition , CNN-LSTM , deep learning
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	02 Oct 2026 00:50
Last Modified:	02 Oct 2026 00:50
URII:	http://shdl.mmu.edu.my/id/eprint/16811

Downloads

Edit (login required)