Hand Sign Language Recognition using Zernike Moments– Based Feature Extraction and Machine Learning Classification

DOI

No DOI

Published In

Volume 5, Issue 2 (2026)

Published Online

May-August 2026

Pages

814-818

Article Type

Original Article

Check for updates

No DOI available for cross-reference

Sign Language Recognition (SLR) Systems bridge the communication gap between speech/hearing-impaired individuals and people. A critical challenge in SLR is creating robust and computationally efficient feature extraction methods that can handle variations in scale, rotation, noise, illumination, and hand shape. This paper proposes a vision-based static sign language recognition framework using Zernike Moments, a set of orthogonal shape descriptors known for their rotation invariance and noise tolerance. The extracted feature vectors are classified using Support Vector Machine (SVM), and Random Forest method. Experimental results demonstrate that the Zernike Moment- based system achieves recognition accuracy of up to 98%, proving its effectiveness for lightweight, real-time SLR applications.