Multi-View Face Detection Based on Improved Adaboost Algorithm
Download as PDF
DOI: 10.23977/csic.2018.0955
Author(s)
Han Zishuo, Wang Chunping
Corresponding Author
Han Zishuo
ABSTRACT
Aiming at the problem of the traditional Adaboost algorithm with fast detection but low accuracy in multi-view face detection, a modified Adaboost algorithm for face detection was proposed in this paper.Firstly,in order to excluding the interference of most non-face regions, the whole image area is detected with skin color segmentation algorithm,then we use cascade structure classifier trained by Adaboost algorithm which addes the extended Haar features to classify the human face candidate regions, thereby the human face can be posited. Experimental results show that the algorithm of this paper can not only effectively improve the human face detection rate and detection speed, but also reduce the false detection rate of multi-view face, in short, the algorithm has a good practical value.
KEYWORDS
Face Detection, Adaboost Algorithm, Skin Color Detection