排序方式: 共有33条查询结果,搜索用时 15 毫秒
1.
This paper presents a Pi-Sigma network to identify first-order Tagaki-Sugeno(T-S) fuzzy inference system and proposes a simplified gradient-based neuro-fuzzy learning algorithm.A comprehensive study on the weak and strong convergence for the learning method is made,which indicates that the sequence of error function goes to a fixed value,and the gradient of the error function goes to zero,respectively. 相似文献
2.
CONVERGENCE OF ONLINE GRADIENT METHOD WITH A PENALTY TERM FOR FEEDFORWARD NEURAL NETWORKS WITH STOCHASTIC INPUTS 总被引:1,自引:0,他引:1
Online gradient algorithm has been widely used as a learning algorithm for feedforward neural network training. In this paper, we prove a weak convergence theorem of an online gradient algorithm with a penalty term, assuming that the training examples are input in a stochastic way. The monotonicity of the error function in the iteration and the boundedness of the weight are both guaranteed. We also present a numerical experiment to support our results. 相似文献
3.
4.
手掌静脉识别典型波长选择 总被引:7,自引:0,他引:7
目前的手掌静脉识别系统均采取主动光源来获取掌脉图像,光源波长的选择直接影响掌脉图像的清晰度与识别性能。典型的掌脉识别成像波长为760,850,890,940 nm,但没有指出哪种波长识别性能最佳。从两个角度解决此问题,从识别特征提取角度,建立了基于Fisher判别率的掌脉成像清晰度模型,对4种波长拍摄的掌脉清晰度进行比较;从特征匹配角度,以3种典型的生物特征识别算法对4种波长拍摄的掌脉图像进行识别性能比较。在包含4种波长共2400幅掌脉图像的自建图库中进行实验,模型选择和典型算法实验结果都表明,850 nm优于其他3种波长。证明了850 nm拍摄的掌脉图像的识别性能最佳。 相似文献
5.
6.
7.
We consider the bifurcation of singular points near a double fold point in Z2 -symmetric nonlinear equations with two parameters,where the linearization has a two dimensional null space spanned by a symmetric null vector and an ami-symmetric null vector. In particular, we show the existence of a turning point path and a pitchfork point path passing ihrough the double fold point and they are the only singular points nearby. Their nondegeneracy is confirmed. A supporting numerical example is also provided. The main tools for our analysis as well as the compulation are some extended systems. 相似文献
8.
解一维二阶椭圆和抛物型微分方程的广义差分法 总被引:8,自引:0,他引:8
本文讨论一维二阶椭圆和抛物方程的广义差分法(一种广义Galerkin方法),试探函数取为Hermite三次元,检验函数取为在小区间中点不连续的分片线性函数,抛物方程的时间方向用Crank-Nicholson型离散,得到了与普通有限元法(试探函数和检验函数都取做Hermite三次元)完全相同的最佳阶误差估计。 相似文献
9.
10.
传统的生物特征识别方法直接将用户的生物特征模板存储于数据库中。由于生物特征具有唯一性和稳定性,一旦被窃取,用户的生物特征将终身不可再用。针对此问题提出了一种存储伪生物特征图像的手掌静脉识别算法,该算法不从外部输入密钥,不存储原始掌脉生物特征模板。在注册阶段,于近红外光下采集用户手掌静脉图像,对图像进行加密形成伪图像,将伪图像存储于数据库中;在认证阶段,将数据库中的伪图像解密后提取特征,与认证阶段采集图像提取的特征进行匹配,给出认证结果。在PolyU图库、CASIA图库和自建图库上进行测试,结果表明:在样本数量为300时,该算法在上述3种图库中的等误率分别为0.4135%、0.5576%、0.4744%,识别时间分别为325.0740,316.0800,322.6530ms。在小范围样本内,所提算法适用于安防、考勤等场合,具有一定的实用价值。 相似文献