THE STUDY AND APPLICATION OF MOMENT INVARIANTS FOR PATTERN RECOGNITION BASED ON GA-BP
The character extraction of moment invariants has been frequently used in pattern recognition, because the moment invariants are invariant to scaling, translation and rotation. When the character extraction of moment invariants is applied in pattern recognition, the values of the moment invariants and similarity measure between recognizing objects and templates must be calculated repeatedly, and then the objects could be recognized according to the values of similarity measure. In this recognizing process, the similarity calculation is not only needed, but also it is hard to determine the threshold of similarity measure. The BP neural network is presented to build a kind of recognition model of moment invariants in this paper and the BP neural network is optimized by Genetic Arithmetic (GA). Taking traffic landmark recognition as an example much better recognition accuracy and speed are not only showed by using GA-BP recognition model of moment invariants, but also the recognition process is simple.
moment invariants pattern recognition Genetic Arithmetic neural network
Qing Liu Jun Wu
School of Automation of Wuhan University of Technology,Wuhan 430063 Hubei China P.R
国际会议
2008年拟人系统国际会议(2008 International Conference on Humanized Systems )(ICHS’08)
北京
英文
2008-10-18（万方平台首次上网日期，不代表论文的发表时间）