中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 西安光学精密机械研究所
  3. 中国科学院西安光学精密机械研究所
  4. 光电测量技术实验室
Sequential three-branch decision method based on Bayesian principle

文献类型：会议论文

作者Zeng, Ke1,2; Ran, Junfeng1
出版日期2021
会议日期2021-05-22
会议地点Kunming, China
关键词Bayesian principle Three-branch decision method Lowest total cost
DOI10.1109/CCDC52312.2021.9601833
页码3902-3907
英文摘要In this paper, the sequential three-branch decision method based on Bayesian principle is applied to face image recognition to realize the minimum risk decision. Adding a pending area to the traditional two-branch decision can minimize the decision cost when the misclassification cost is imbalanced and the feature information of the sample is temporarily insufficient. In addition, the acquisition of sample feature information is not unlimited. Although each feature information obtained can reduce the cost of misclassification of decision, it is also necessary to consider the increased test cost of each feature information obtained. This paper used the Bayesian principle to realize the dynamic balance of them to achieve the minimum total decision cost (misclassification cost and test cost). In this paper, 2DPCA was used to obtain the granular features of face images, and experiments were conducted on face databases such as AR and PIE to verify the effectiveness of the sequential three-branch decision method based on Bayesian principle. © 2021 IEEE.
产权排序1
会议录Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021
会议录出版者Institute of Electrical and Electronics Engineers Inc.
语种英语
ISBN号9781665440899
源URL[http://ir.opt.ac.cn/handle/181661/95757]  
专题西安光学精密机械研究所_光电测量技术实验室
作者单位1.Xi'An Institute of Optics and Precision Mechanics of CAS, Xi'an; 710119, China
2.University of Chinese Academy of Sciences, Beijing; 100049, China;
推荐引用方式
GB/T 7714		 Zeng, Ke,Ran, Junfeng. Sequential three-branch decision method based on Bayesian principle[C]. 见:. Kunming, China. 2021-05-22.

入库方式： OAI收割

来源：西安光学精密机械研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。