Design and implementation of adaptive learning based on ART neural network

In this paper, we present a new method that it has provided the learning function to solve the problem of diagnostic images recognition effectively. The human brain's characteristics are studied and referenced by this research, for examples, in the first place, the human brain's study is autonomous, it can study and identify an objective in a complex, unsteady, and a interference environment, what is more, it study on its own under a majority of situation, and synchronism is in progress between the study and practice, in addition, the human brain's memory is provide the characteristic of self-organized obviously. Moreover, this technology uses the characteristic of the plasticity and connectible. The ART neural network architecture is used and the rules are improved, the paper provided an algorithm with a function of class learning such as self-adaptive, self-steady and self-study, etc. Finally, to suit the multi-channel input signals of more objects, and to solve the complex diagnostic images recognition effective, a method of self-adaptive and class learning is presented.