Many common foundations exist between neural networks and fuzzy inference systems in terms of their mathematical models and system structures. This paper explores such a rich synergy and uses it to form the basis for a unifying framework under which fuzzy logic processing and neural networks may be integrated to achieve more robust information processing. It in turn leads to a family of hierarchical fuzzy neural networks (FNN's) which incorporate an adaptive and modular design of neural networks into the basic fuzzy logic systems. Several important models which are critical to the development of the hierarchical FNN family are studied carefully so as to gain a better understanding of how the FNN's could benefit from the symbiotic marriage of the learning techniques of neural networks and the inference structure of fuzzy logic systems. Specifically, we demonstrate how existing unsupervised [e.g., competition-based learning, expectation-maximization (EM) algorithm] and supervised (e.g., reinforced/antireinforced learning) training strategies can be an integral part of a fuzzy processing framework. In addition, for robust processing, hierarchical structures involving both expert (rule) modules and class modules are incorporated into the FNN's. Also presented are some promising application examples for biometric authentication, medical image processing, video segmentation, object recognition/detection, and multimedia content-based retrieval.
All Science Journal Classification (ASJC) codes
- Electrical and Electronic Engineering