Study of the cost-sensitive AdaBoost face detection algorithm
Dingli Song, Bingru Yang, Fuxing Yu · 2010
It is the mainstream method that in human face detection and recognition with AdaBoost as the representative based on statistical learning method. Detection rates have reached a high level, and can achieve real-time detection. However, AdaBoost algorithm treats equally for different categories, there is no distinction between the cost of the different error categories. This paper presents a new cost-sensitive AdaBoost-based face detection algorithm, ensuring the detection rate and speed, effectively reducing the false detection rate, and improving the detection accuracy.