Human Gender and Age Estimation on Real-time Video
Implemented Biologically Inspired Features algorithm for feature extraction, Kernel Partial Least Square Regression for age and gender prediction, and Viola-Jones algorithm for face detection from real-time video in MATLAB using FG-NET dataset. Best MAE achieved: 2.43 years and gender prediction power: 85.36%
Also K Nearest Neighbour algorithm is used for age group classification (very young, young, middle, old age).