Refining Pre-image via Error Compensation for KPCA-based Pattern Denoising

被引：0
作者：
Li, Jianwu [1 ]
Tu, Qiang [1 ]
Yan, Ziye [1 ]
机构：
[1] Beijing Inst Technol, Sch Comp Sci & Technol, Beijing Key Lab Intelligent Informat Technol, Beijing 100081, Peoples R China
来源：
2016 23RD INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR) | 2016年
关键词：
kernel principal component analysis (KPCA); pre-image; systematic error; error compensation;
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Finding pre-image is crucial for kernel principal component analysis (KPCA) based pattern de-noising. This paper proposes to learn the systematic error of some classical methods of pre-image finding, and to refine the obtained pre-image via error compensation. Experiments based on simulated data as well as real-world data demonstrate that the proposed approach can improve effectively the results from two classical pre-image methods: gradient decent and distance constraint.
引用
收藏
页码：414 / 419
页数：6
相关论文
共 13 条
  • [11] Input space versus feature space in kernel-based methods
    Schölkopf, B
    Mika, S
    Burges, CJC
    Knirsch, P
    Müller, KR
    Rätsch, G
    Smola, AJ
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 1999, 10 (05): : 1000 - 1017
  • [12] Wei-Shi Zheng, 2006, 2006 18th International Conference on Pattern Recognition
  • [13] Penalized Preimage Learning in Kernel Principal Component Analysis
    Zheng, Wei-Shi
    Lai, JianHuang
    Yuen, Pong C.
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 2010, 21 (04): : 551 - 570
12