Data Fusion and Confidence in Image Feature Detection and Mahalanobis Distance
图像特征检测和马氏距离中的数据融合与置信度

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.