论文部分内容阅读
文献[2][3]以谱半径的形式给出了循环侧抑制网络的稳定性条件。在实际图像处理中,应用此条件判断网络的稳定性并不方便。本文在此结论的基础上,通过理论证明,给出了在有限抑制野的情况下,侧抑制系数应该满足的显式条件,易于进行网络稳定性判断。最后,应用选定的侧抑制系数处理标准的Lenna图像,结果显示了算法在图像预处理中增强反差,突出边框的作用。
In [2] and [3], the stability conditions of the network with cyclic suppression are given in the form of spectral radii. In actual image processing, it is inconvenient to use this condition to judge the stability of the network. On the basis of this conclusion, the paper proves that the explicit condition that the side suppression coefficient should satisfy in the case of limited field suppression is given, which makes it easy to judge the network stability. Finally, the standard Lenna image is processed with the selected side suppression coefficients. The results show that the algorithm enhances contrast and emphasizes the border in the image preprocessing.