Robust Designs for Templates of Directional Extraction Cellular Neural Network with Application

The cellular neural/nonlinear network (CNN) is a powerful tool for image and video signal processing, robotic and biological visions. In this paper, the selected objects extraction (SOE) CNN was generated to directional extraction (DE) CNN which enhance the capabilities of CNNs and improve their efficiency. Based on analytical approach, a theorem of designing robust templates for DE CNNs was established, which provides parameter inequalities to determine parameter intervals for implementing the corresponding functions. Several examples are provided to illustrate the effectiveness of the theorem for extracting selected objects directionally in binary images.