Adaptive Fuzzy Evidential Reasoning Data Fusion Scheme and its Application to Brain Tissue Segmentation

This paper presents an adaptive fuzzy evidential reasoning approach for multi source based data fusion. A novel fuzzy evidence structure model is proposed under the assumption that each information source provides two types of evidence: probabilistic evidence (in terms of posteriori probabilities) and fuzzy evidence (in terms of fuzzy rules). A new information measure, called hybrid entropy, is defined for evaluating the overall uncertainty contained in a fuzzy evidence structure. For adaptive reasoning, two discounting strategies are introduced to handle potential conflicts between probabilistic evidence and fuzzy pieces of evidence. First, a local discounting scheme is introduced to account for relationship between the two types of evidence. Subsequently, a global discounting scheme is introduced to make use of source quality so as to deal with conflict among the information sources. To demonstrate the effectiveness of the proposed approach we apply it to automated brain tissue segmentation based on multi-modality MR images. It is concluded that the proposed approach outperforms commonly used techniques such as: K-means clustering, majority voting, fuzzy set operators based and Baysian based approaches.