摘要
针对体育比赛期间运动员的检测识别并显示对应统计信息问题,提出了一种运动员人脸检测识别方法 .主要分四个步骤:第一步,利用一阶卡尔曼滤波器跟踪被检测目标;第二步,使用AdaBoost和Haarlike特征检测进行特征选择和分类;第三步,利用推进方法进行人脸检测;第四步,利用LDA初始化的AdaBoost算法识别人脸.用数码相机采集412张不同图像进行实验.实验结果表明,本文提出的方法在大部分情况下都能获得最高的球员检测精度和人脸识别精度.
For the problems of automatic detection,recognition of athletes during play,followed by display of personal information of athletes,we proposed a new face detection and recognition method for athletes.In the first step,each player in the image is detected by the first order Kalman filter.In the second step,AdaBoost algorithm with Haar like features is used for both feature selection and classification.In the third step,the propulsion method is used for face detection.In the fourth step,we utilized AdaBoost algorithm with LDA as a weak learner for feature selection in LDA subspace.Detailed experiments are performed using412diverse images taken by a digital camera during baseball match.The result shows that the face detection precision and the face recognition accuracy are high in most situations.
作者
张馨娇
李建萍
ZHANG Xinjiao;LI Jianping(Physical Education and Research Department, Shangluo University, Shangluo 726000;School of Computer and Communication, Lanzhou University of Technology, Lanzhou 730050 China)
出处
《湘潭大学自然科学学报》
北大核心
2017年第4期95-98,共4页
Natural Science Journal of Xiangtan University
基金
国家自然科学基金项目(11471007)
陕西省教育厅专项科研项目(2017:17JK0226)