基于模糊聚类的语音识别

指出模糊聚类方法影响语间识别的正确率 ,提出用基于分解构造的模糊聚类方法设立语音模式的参考向量 ,说明该方法能提高语章识别的正确率。     

A new nonlocal model for the restoration of textured images

In this paper, we focus on the mathematical and numerical study of a new nonlocal reaction-diffusion system for image denoising. This model is motivated by involving the decomposition approach of $H^{-1}$ norm suggested by Meyer [25] which is more appropriate to represent the oscillatory patterns and small details in the textured image. Based on Schaeffer''s fixed point theorem, we prove the existence and uniqueness of solution of the proposed model. To illustrate the efficiency and effectiveness of our model, we test the denoising experimental results as well we compare with some existing models in the literature.     

Recognizing Balanceable Matrices

A 0/±1 matrix is balanced if it does not contain a square submatrix with exactly two nonzero entries per row and per column in which the sum of all entries is 2 modulo 4. A 0/1 matrix is balanceable if its nonzero entries can be signed ±1 so that the resulting matrix is balanced. A signing algorithm due to Camion shows that the problems of recognizing balanced 0/±1 matrices and balanceable 0/1 matrices are equivalent. Conforti, Cornuéjols, Kapoor and Vušković gave an algorithm to test if a 0/±1 matrix is balanced. Truemper has characterized balanceable 0/1 matrices in terms of forbidden submatrices. In this paper we give an algorithm that explicitly finds one of these forbidden submatrices or shows that none exists. Received: October 2004     

Variables separated equations: Strikingly different roles for the Branch Cycle Lemma and the finite simple group classification

Davenport’s Problem asks:What can we expect of two polynomials,over Z,with the same ranges on almost all residue class fields? This stood out among many separated variable problems posed by Davenport,Lewis and Schinzel.By bounding the degrees,but expanding the maps and variables in Davenport’s Problem,Galois stratification enhanced the separated variable theme,solving an Ax and Kochen problem from their Artin Conjecture work.Denef and Loeser applied this to add Chow motive coefficients to previously introduced zeta functions on a diophantine statement.By restricting the variables,but leaving the degrees unbounded,we found the striking distinction between Davenport’s problem over Q,solved by applying the Branch Cycle Lemma,and its generalization over any number field,solved by using the simple group classification.This encouraged Thompson to formulate the genus 0 problem on rational function monodromy groups.Guralnick and Thompson led its solution in stages.We look at two developments since the solution of Davenport’s problem.Stemming from MacCluer’s 1967 thesis,identifying a general class of problems,including Davenport’s,as monodromy precise.R(iemann)E(xistence)T(heorem)’s role as a converse to problems generalizing Davenport’s,and Schinzel’s (on reducibility).We use these to consider:Going beyond the simple group classification to handle imprimitive groups,and what is the role of covers and correspondences in going from algebraic equations to zeta functions with Chow motive coefficients.     

Vague集信息熵测量及其应用

引入了区别于现有文献的Vague集信息熵和Vague集的关联熵的概念,给出了一种改进的测量方法,并讨论了它们之间的关系。进而,我们揭示了Vague集的熵和Fuzzy集的熵之间的关系,并分析了本文所定义熵的意义。最后,讨论了这种关联熵在模糊识别和医疗诊断上的应用。     

Post-primary students’ images of mathematics: findings from a survey of Irish ordinary level mathematics students

A questionnaire survey was carried out as part of a PhD research study to investigate the image of mathematics held by post-primary students in Ireland. The study focused on students in fifth year of post-primary education studying ordinary level mathematics for the Irish Leaving Certificate examination – the final examination for students in second-level or post-primary education. At the time this study was conducted, ordinary level mathematics students constituted approximately 72% of Leaving Certificate students. Students were aged between 15 and 18 years. A definition for ‘image of mathematics’ was adapted from Lim and Wilson, with image of mathematics hypothesized as comprising attitudes, beliefs, self-concept, motivation, emotions and past experiences of mathematics. A questionnaire was composed incorporating 84 fixed-response items chosen from eight pre-established scales by Aiken, Fennema and Sherman, Gourgey and Schoenfeld. This paper focuses on the findings from the questionnaire survey. Students’ images of mathematics are compared with regard to gender, type of post-primary school attended and prior mathematical achievement.     

混合优化RBF-BP网络的板形缺陷识别

针对传统板形模式识别方法存在精度低、鲁棒性弱的问题,提出了一种混合优化RBF-BP组合神经网络板形模式识别方法。首先利用自组织映射网络(SOM)对样本聚类,利用聚类后的网络拓扑结构确定RBF的中心,并计算RBF的宽度,克服了传统聚类算法随机选取中心导致聚类结果不稳定的问题。然后利用遗传算法(GA)良好的全局搜索能力优化整个网络的权值。RBF-BP组合神经网络是由一个RBF子网和一BP子网串联构成的,该网络同时具备BP神经网络能较好地预测未知样本的能力以及RBF神经网络的逼近速度快的优点。并以某900HC可逆冷轧机板形识别为应用背景,在MATLAB2010a环境下进行仿真实验,结果表明混合优化RBF-BP组合神经网络的板形模式识别方法能够识别出常见的板形缺陷,提高了板形缺陷识别精度并具有较好的鲁棒性,可以满足板带轧机高精度的板形控制要求。     

N指标d维广义Brownian Sheet的像集及图集的Hausdorff测度

本文得到了Ｎ指标ｄ维广义Ｂｒｏｗｎｉａｎ Ｓｈｅｅｔ的像集及图集的Ｈａｕｓｄｏｒｆｆ测度的上、下界．     

基于方向信息测度的图像边缘检测

边缘检测是图像处理中—个重要的研究课题．针对传统图像边缘检测算法对噪声敏感的问题,本文在分析图像像素灰度信息的基础上,建立了—个改进的确定方向信息测度的方法,并利用震动滤波对边缘检测图像进行增强,该方法在滤除噪声的同时,能有效地保留图像的基本目标信息,正确提取图像的边缘．     

基于小波与齐次Besov空间的图像分割算法

从具有全局最优解的几何活动轮廓方法出发,分别提出了两种基于齐次Besov窄间与小波变换的图像分割算法,并给出了解的存在性证明.数值求解利用小波软阈值以及分裂Bregman方法,能够有效提高计算效率.由于小波变换具有多分辨特性,对于包含较多细节信息的图像,采用新算法能够得到更好的分割效果.数值实验表明采用新算法能够获得较...