摘要
提出了一种结合肤色模型和Adaboost算法的人脸检测方法.Adaboost是一种分类器算法,把弱分类器提升为强分类器,具有很高的检测率,但其误检率也比较高.人脸肤色检测由于受光照等条件的影响,其检测率比较低,但是其误检率也比较低.结合这两种方法进行检测,能够优势互补,改善算法性能.实验证明该算法具有较好的检测效果.
A novel face detection method combined skin color model and adaboost algorithm is proposed. Adaboost algorithm boosting weak classier to strong classier is a kind of classifier algorithm with high detection rate, but it also has high false positive rate. Because of light conditions, color skin detection method has low defection rate, while it＇s false positive is also low. This new method takes advantage of the two algorithms, and thus its detection ability is enhanced. The experimental result proves the method has good performance.
出处
《苏州大学学报（自然科学版）》
CAS
2009年第3期63-67,共5页
Journal of Soochow University(Natural Science Edition)
基金
国家自然科学基金资助项目(60678051)
