Integrated Boolean neural networks for feature extraction and image classification

The recognition of two-dimensional images is important in many diverse areas of application ranging from character recognition to automatic industrial inspection and medical image processing, and many possible approaches to system design for such applications exist. Because of the inherently parallel nature of the algorithms conventionally adopted in classification processing, many recent attempts to improve the flexibility and effectiveness of classifier design have focused on the implementation level, exploiting parallel architectures to provide a judicious match to task specification. There is, however, considerable interest in developing new parallel architectures based on neural network designs which can be applied to practical classification tasks