Robust Face Detection Using Hybrid Skin Color Matching under Different Illuminations

Abstract

Recently, human skin color property incorporation is considered a way to improve face detection. There are different skin color matching techniques and each individual one is found suitable for a specific illumination condition. In this study, a robust face detection is developed considering a hybrid skin color detection method which combines complementary strength of different color detection methods. The proposed hybrid skin color detection method is tested with the most popular Haar Feature Based Face Detection (HFFD) method and identified performance on a large number of images from a benchmark face data set. From the experimental results and analysis, it is revealed that proposed face detection system improves the performance of HFFD significantly.