图像特征检测和马氏距离中的数据融合与置信度  

Data Fusion and Confidence in Image Feature Detection and Mahalanobis Distance

作　　者：朱秋煜[1] 王朔中[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《电子与信息学报》2008年第3期534-538,共5页Journal of Electronics & Information Technology

基　　金：上海市重点学科建设项目(T0102)资助课题

摘　　要：该文在单源图像的模式特征检测中应用数据融合,用统计方法对同一特征的多算法结果进行融合,论证算法的合理性和融合结果的有效性,定义融合数据的置信度以反映其可靠程度。在此基础上推导了包含置信度的马氏距离度量公式。以人脸检索及其中的眼角检测为例,进行了多算法结果的融合并讨论其置信度,说明该文提出的方法能有效降低模式分类时错误检测参数的不利影响,提高模式识别率。Using a statistical approach, data fusion is performed in image feature detection with which information about the same feature obtained by multiple methods can be integrated. Validity of the described fusion scheme and properties of the fused data are discussed. A confidence measure is defined and applied to evaluate credibility of the results. Taking into account the data fusion and confidence measure, a Mahalanobis distance is derived. The technique is applied to face retrieval and its canthus detection. Experimental results show that the proposed approach can reduce adverse effects of feature detection errors and enhance the pattern recognition rate.

关 键 词：特征检测 数据融合 置信度 马氏距离 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]