基于肤色和Adaboost算法的人脸检测研究

Face Detection Based on Skin Color and Adaboost Arithmetic

人脸检测是计算机视觉和人工智能领域中的一项富有挑战性的工作,在虚拟现实、人机交互等很多领域都有广泛的应用。研究了基于Adaboost的人脸检测,并提出了肤色与Adaboost算法相结合的人脸检测方法。对输入的彩色图像进行从RGB空间到YCrCb空间的转换,再结合形态学等方法进行区域肤色分割,排除背景干扰,然后用Adaboost算法对可能区域进行检测,得到人脸位置。实验表明,该方法有较高的准确性和鲁棒性,可以得到满意的检测效果。

Face detection is a challenge in computer vision and artificial intelligence area, and applies in birtual reality, human- computer interaction techniques and so on.Studies the face detection based on Adaboast arithmetic,and presents a face detection method based on skin color and Adaboost arithmetic. First transforms the input color images form RGB space to YCrCb space,then uses morphologic methods to make area skin segmentation,at last uses Adaboost arithmetic to detect the face in the possible area, and gets the face position. The experiments show that this method has high accuracy and robusmess, can get satisfied detect results.