摘要
提出了一种基于小波神经网络的掌纹识别方法。首先对掌纹图像经过预处理得到掌纹的感兴趣区域(ROI),然后利用小波包分解的方法对该区域进行掌纹特征的提取,再利用RBF网络的容错能力和较快的收敛性对掌纹图像加以识别。针对香港理工大学掌纹数据库进行了实验,实验结果证明,本算法可以达到很好的识别效果,为掌握识别提供了一种新途径。
A palmprint recognition method based on wavelet and neural network was proposed.Firstly,palmprint image preprocessing was used to obtain the region of interest.Then wavelet packet decomposition was applied to extract the palmprint feature.At last,the palmprint image was recognized by RBF network with good fault tolerance ability and fast convergence.The experiments were implemented on the Hong Kong Polytechnic University（PolyU） Palmprint Database.The results improved that this algorithm could achieve a good recognition effect and provided a new approach for palmprint recognition.
出处
《齐齐哈尔大学学报（自然科学版）》
2010年第3期5-8,共4页
Journal of Qiqihar University(Natural Science Edition)
关键词
感兴趣区域
小波包分解
RBF神经网络
region of inerest
wavelet packet decomposition
RBF neural network
