Robust Designs of Selected Objects Extraction CNN
Fangyue Chen, Lin Chen, Weifeng Jin · 2009
The cellular neural/nonlinear network (CNN) is a powerful tool for image and video signal processing, robotic and biological visions. In this paper, the robust CNN template for extracting the selected objects in binary images is designed, and the parameter inequalities for determining parameter intervals for implementing the corresponding tasks are provided. The selected objects extraction CNN derived in this paper can successfully extract marked objects with the patterns connecting each other via "edges" or corners. In addition, two examples are provided to illustrate the effectiveness of the selected objects extraction CNN.