Ethnicity Classification using Deep Learning
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Malaysia Sarawak
Abstract
Description
Classifying ethnicity is an important field of study with various practical applications in identity verification, personalized user experiences, demographic studies, and bias mitigation in artificial intelligence (AI) systems. As facial recognition technology becomes increasingly integrated into societal functions, ensuring accurate and fair ethnicity classification is essential to prevent the perpetuation of biases and inaccuracies. This study explores ethnicity classification using deep learning, utilizing a Convolutional Neural Network (CNN) built upon the VGG-Face architecture, which is widely recognized for its strong performance in facial recognition applications. Despite advancements in ethnicity classification, existing methods often lack diverse datasets and struggle with generalizability, limiting their effectiveness in multicultural societies like Malaysia. To address these gaps, this research aims to develop and evaluate a deep learning model capable of accurately classifying ethnicities in Malaysia, focusing on major ethnic groups such as Malay, Chinese, and Indian. A custom dataset representing Malaysia's ethnic diversity was curated to ensure inclusivity and fairness, overcoming the limitations of global datasets that lack representation of local ethnic characteristics. These findings highlight the capabilities of deep learning to improve demographic data collection and analysis. By enabling more equitable resource allocation and better-tailored services, this study adds value to the domain of demographic analysis by providing a scalable, accurate, and ethical solution for ethnicity classification. Its applications extend to governance, healthcare, and social services, contributing to efforts that promote inclusivity and improve service delivery in Malaysia’s diverse population.