摘要
ASM主动形状模型是一种基于点分布的模型。在该模型中,外观相似的物体,可以通过一定数量的关键点坐标来表示出位置和相对应的顺序。在常规的基于ASM模型的人脸跟踪技术中,会发生拟合点漂移的现象,造成定位的不准确。文中提出一种结合了Adaboost人脸跟踪和传统的ASM模型的算法,能够有效地减少在传统ASM模型中出现的跟踪漂移现象。
ASM （Active Shape Model） is a model based on points distribution. In this model, the objects which have similar appearance can be expressed by a number of feature points in a specific order. In regular ASM model, sometimes a point drift phenomenon would happen, which causes incorrect feature positioning. This article presents a algorithm which combines Adaboost face tracking algorithm and regular ASM model algorithm and this new algorithm can effectively reduce the tracking drift phenomenon.
出处
《信息技术》
2015年第7期88-89,94,共3页
Information Technology
