| 本文已被：浏览 5851次 下载 7807次 |
码上扫一扫！ |
|
|
| 对数-指数形态学联想记忆 |
|
冯乃勤1,2, 田勇3, 王鲜芳1, 宋黎明1, 范海菊1, 王双喜4
|
|
1.河南师范大学 计算机与信息工程学院, 河南 新乡 453007;2.郑州工业应用技术学院 信息工程学院, 河南 郑州 451100;3.河南机电职业学院 信息工程系, 河南 郑州 451191;4.商丘学院 计算机科学与技术学院, 河南 商丘 476000
|
|
| 摘要: |
| 利用对数和指数算子构建了一种新的形态学联想记忆方法,简称LEMAM.理论分析表明:自联想LEMAM(简称ALEMAM)具有无限存储能力、一步回忆记忆、一定的抵抗腐蚀噪声或膨胀噪声的能力,在输入完全或在一定的噪声范围内,能够保证完全回忆记忆;异联想LEMAM(简称HLEMAM)在输入完全情况下,不能保证完全回忆记忆,但当满足一定条件时,也能够达到完美联想记忆.对比实验结果表明:在一些情况下,LEMAM能够取得较好的联想记忆效果.总体来说,LEMAM丰富了形态学联想记忆的理论和实践,可以作为一种神经计算模型加以研究和利用. |
| 关键词: 对数 指数 形态学联想记忆 自联想记忆 异联想记忆 完全回忆记忆 |
| DOI：10.13328/j.cnki.jos.004620 |
| 分类号: |
| 基金项目:国家自然科学基金(61173071); 河南省高校创新人才支持计划(2012HASTIT011) |
|
| Logarithmic and Exponential Morphological Associative Memories |
|
FENG Nai-Qin1,2, TIAN Yong3, WANG Xian-Fang1, SONG Li-Ming1, FAN Hai-Ju1, WANG Shuang-Xi4
|
|
1.College of Computer and Information Engineering, He'nan Normal University, Xinxiang 453007, China;2.School of Information Engineering, Zhengzhou University of Industrial Technology, Zhengzhou 451100, China;3.Department of Information Engineering, He'nan Mechanical and Electrical Vocational College, Zhengzhou 451191, China;4.College of Computer Science and Technology, Shangqiu University, Shangqiu 476000, China
|
| Abstract: |
| A novel morphological associative memory method, abbreviated as LEMAM, is constructed by using logarithmic operator and exponential operator. The theoretical analysis shows that auto LEMAM (abbreviated as ALEMAM), which has unlimited storage capacity, one step recall, and a certain ability of resisting erosive noise or dilative noise, can ensure perfect recall memory for either perfect inputs or a certain range of noise. Hetero LEMAM (abbreviated as HLEMAM) does not guarantee perfect recall, even without any input noise. However, when meeting certain conditions, HLEMAM can also achieve perfect recall. HLEMAM contrast experiments show that, in some cases, LEMAM can produce better result. On balance, LEMAM enriches the theory and practice of morphological associative memories, and can serve as a kind of new neural computational model for research and application. |
| Key words: logarithm exponent morphological associative memories autoassociative memories heteroassociative memories perfectrecall |
