CASIA OpenIR  > 毕业生  > 硕士学位论文
集成型模型识别系统的理论与应用
汪力新
Subtype工学硕士
Thesis Advisor戴汝为
1998-05-01
Degree Grantor中国科学院自动化研究所
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword综合集成 人机结合 集成型模式识别系统 监督学习 反馈 Metasynthesis Human-machine Integration Integrated Pattern Recognition System Supervised Learning Feedback
Abstract本文在人机结合的综合集成思想的指导下,对单分类器进行集成,以充分 发挥人的指导作用。首先提出了含子类的竞争监督学习算法,然后受控制系统 反馈思想的启发,在集成阶段采取闭环结构,使其成为一个动力学系统。论文 详细分析了它的稳定性质,并且在考虑稳定条件的基础上给出相应的监督学习 算法。相应的手写数字识别实验表明所提模型的有效性。 此外,论文还介绍了作者在不变矩以及微分滤波器等方面的工作
Other AbstractIn this dissertation, enlightened by the idea of metasynthesis, an Integration method is proposed to combine the human intelligence with the capability of computers. We first introduce the Competitive Supervised Learning algorithm with sub-classes. Then we implement a closed loop in the integrated neural network so that it becomes a dynamic system similar to those in cybernetics. The dissertation analyzes its dynamic property and gives the corresponding supervised learning algorithm under the proposed stable conditions. The experiments on the totally unconstrained handwritten numeral verified its superiority. On the other hand, this dissertation also gives the work on moment invariants and differential filters, respectively.
shelfnumXWLW469
Other Identifier469
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/7220
Collection毕业生_硕士学位论文
Recommended Citation
GB/T 7714
汪力新. 集成型模型识别系统的理论与应用[D]. 中国科学院自动化研究所. 中国科学院自动化研究所,1998.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[汪力新]'s Articles
Baidu academic
Similar articles in Baidu academic
[汪力新]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[汪力新]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.