Implementing morphological operations using programmable neural networks

Neural networks have been studied for decades to achieve human-like performances. There has been a recent resurgence in the field of neural networks caused by new topologies and algorithms, analog VLSI implementation techniques, and the belief that massive parallelism is essential for high performance image and speech recognition. This paper presents a novel idea of implementing image morphological operations using programmable neural networks. The architecture has the optional programmable logic/analogy framework, hence, it can handle a variety of binary and gray-scale processings and avoid some of the limitations of threshold logic networks. An example of applying this network to illustrate the activation of neocognitron for visual pattern recognition is also provided.