Face detection using combination of skin color pixel detection and Viola-Jones face detector

Abstract

This paper presents a human face detection scheme by combining a novel hybrid color models and Viola-Jones face detector. A hybrid skin color model RGB-CbCrCg was proposed for classifying skin and non-skin pixels. The extraction of skin region is carried out using a set of bounding rules optimized employing multi-objective differential evolution method. Afterward the segmented face regions are identified using Viola-Jones algorithm built in MATLAB Computer Vision System Toolbox™. The performances of skin and face detection were evaluated using the ECU face and skin database. Experimental results show that the proposed scheme outperforms the other skin color space methods.