Research on the Dual Kernel Based on the Unified Framework of Morphological Associative Memories

摘要：

The Morphological Associative Memories (MAM) are a class of extremely new artificial neural networks. Although MAM has many advantages, it can not have a good performance of tolerating random noises and mixture noises. In order to solve this problem, some people have presented the kernel method, but there is still room to research. So we research on the kernel from another angle. On the one hand we present the concept of the dual kernel based on the unified Framework of Morphological Associative Memories, on the other hand, we research on its properties in-depth. At last, through the computing and simulation experiments of true-color images, we confirm that the dual kernel has the validity of tolerating random noise and mixture noise.

关键词： component unified framework of morphological associative memories the dual kernel the recall performance

作者: Naiqin Feng Shuangxi Wang Jiucheng Xu Shanxia Wang Yong Tian

作者单位: College of Computer and Information Technology Henan Normal University, 453007 Xinxiang, China

会议类型: 国际会议

会议名称: The 2010 International Conference on Computer Application and System Modeling(2010计算机应用与系统建模国际会议 ICCASM 2010)

会议地点: 太原

会议语种:英文

页码: 643-648

在线出版日期: 2010-10-22（万方平台首次上网日期，不代表论文的发表时间）