An optimized algorithm for recognition of complex patterns based on artificial neural network

The paper presents an optimized algorithm for the recognition of complex and noisy patterns. The system, based on the backpropagation neural network (NN), can be used in numerous applications. One of them is pattern recognition in biomedical signals, used for example in wireless body area networks. In the paper the efficiency of the system has been verified with data set composed of letters with not typical shapes, additionally covered by a noise. The subject of the optimization process was the network sizes i.e. the number of neurons and the number of layers, as well as different learning parameters. The proposed system has been written in Java language. It can be used as a model of a real system that in the next step will be realized in hardware, for example, as an ASIC. The system enables creation of NNs with different parameters, and a full learning process that also includes the testing phase. It is equipped with the editor of both the raster and matrix patterns.