论文信息 - Modified Back Propagation Algorithm for CNN Template Design

Modified Back Propagation Algorithm for CNN Template Design

In the previous study, we have proposed a template design method of cellular neural networks with back propagation algorithm. In that method, template learns by using the average error which corresponds to the difference between the output image and the desired image. In this study, we modify the back propagation algorithm for cellular neural networks template design. We inspect the performance of this learning algorithm for two-dimensional binary image.

Yoshifumi Nishio | Yoko Uwate | Masashi Nakagawa

[1] Y. Nishio,et al. CNN template design using back propagation algorithm , 2010, 2010 12th International Workshop on Cellular Nanoscale Networks and their Applications (CNNA 2010).

[2] Lin-Bao Yang,et al. Cellular neural networks: theory , 1988 .

[3] Uwate Yoko,et al. Template Design of CNN by BP with Annealing Noise , 2010 .

[4] Geoffrey E. Hinton,et al. Learning internal representations by error propagation , 1986 .