3-dimensional pattern recognition using weightless neural networks (Aleksander's model)

A system composed of a weightless neural network (Aleksander's model) is described. This system performs 2D and 3D recognition of patterns which are drawn with five different gray tones or colors. The system is capable of recognizing from 2 to 15 different pattern classes. The authors describe the structure and behavior of this system during the learning and recognition phases.<
>