Advance Facial Analysis: A Hybrid Approach with CNNs, SVMs, and Random Forest for Age and Gender Estimation
Kulvinder Singh, Balkar Singh · 2024
In an era marked by the widespread adoption of Artificial Intelligence (AI) and advanced computer vision technologies, the analysis of facial attributes has become increasingly important across various domains such as access control, video surveillance, and personalized services. A single facial image contains a wealth of demographic information including age, gender, race, and emotional expression. The paper focuses on a unique approach for facial image analysis that combines Convolutional Neural Networks (CNNs) for detailed feature extraction, traditional models like Support Vector Machine for age prediction, and Random Forest also for gender classification. The proposed approach not only addresses existing limitations but also enhances predictive accuracy. The effectiveness of our method was extensively evaluated using the Adience Benchmark dataset, producing outstanding results in both age prediction and gender classification tasks. Through thorough assessments across multiple datasets and evaluation techniques, the proposed solution showcases improved performance as compared to current state of the art methods, affirming its efficacy and reliability.