Research on an improved algorithm of face detection based on skin color features and cascaded Ada Boost

摘要

The paper presents an improved algorithm in face detection that combines the YcbCr color space to detect the skin region and the AdaBoost algorithm to precise detection. The Ada Boost algorithm is one technique among the most successful algorithms in face detection, but in view of its time costs and high false acceptance rate, the paper proposes a two-stage detection method based on skin color segmentation which has relatively high speed of operation at the expense of accuracy and the cascaded AdaBoost algorithm which has high accuracy in face detection but time consuming. Here the modified algorithm research uses the open source platform Open Cv to simulate. The result shows the algorithm is more capable for real-time face detection.