共查询到20条相似文献,搜索用时 15 毫秒
1.
分析Zernike矩人耳特征提取和非负矩阵分解(NMF)人耳特征提取的利弊。将线性判别分析的思想融入到NMF算法中,对传统的NMF方法进行改进。介绍一种融合特征人耳识别方法:将Zernike矩和传统非负矩阵分解融合提取人耳特征,得到一个分类能力更强的人耳特征矩阵,并采用BP神经网络进行分类识别,实验结果表明,应用融合特征方法提取人耳图像特征,可以提高识别效果。 相似文献
2.
3.
主成分分析(PCA)和线性判别分析(LDA)是基于全局结构的特征提取方法,局部保持投影(LPP)和正交拉普拉斯脸(OLF)是基于局部结构的特征提取方法,全局结构特征的弱点是忽略了局部结构特征,局部结构特征的弱点是忽略了整体结构.基于此提出了一种全局与局部结构图像特征融合(GLSF)的提取方法,将PCA和LDA的提取结果融合到LPP中,既描述了全局结构,又考虑了局部结构.在ORL及Yale上的实验结果表明,GLSF方法比PCA,LDA,LPP,OLF等方法具有更高的识别率. 相似文献
4.
5.
Conventional hyperspectral image-based automatic target recognition (ATR) systems project high-dimensional reflectance signatures onto a lower dimensional subspace using techniques such as principal components analysis (PCA), Fisher's linear discriminant analysis (LDA), and stepwise LDA. Typically, these feature space projections are suboptimal. In a typical hyperspectral ATR setup, the number of training signatures (ground truth) is often less than the dimensionality of the signatures. Standard dimensionality reduction tools such as LDA and PCA cannot be applied in such situations. In this paper, we present a divide-and-conquer approach that addresses this problem for robust ATR. We partition the hyperspectral space into contiguous subspaces based on the optimization of a performance metric. We then make local classification decisions in every subspace using a multiclassifier system and employ a decision fusion system for making the final decision on the class label. In this work, we propose a metric that incorporates higher order statistical information for accurate partitioning of the hyperspectral space. We also propose an adaptive weight assignment method in the decision fusion process based on the strengths (as measured by the training accuracies) of individual classifiers that made the local decisions. The proposed methods are tested using hyperspectral data with known ground truth, such that the efficacy can be quantitatively measured in terms of target recognition accuracies. The proposed system was found to significantly outperform conventional approaches. For example, under moderate pixel mixing, the proposed approach resulted in classification accuracies around 90%, where traditional feature fusion resulted in accuracies around 65%. 相似文献
6.
决策层信息融合的神经网络模型与算法研究 总被引:8,自引:0,他引:8
本文对信息融合问题中决策层融合方法进行了分析与比较,提出了一种新的决策层信息融合算法,即改进型ART2神经网络融合算法,该融合算法在综合大脑对多源信息融合的特点和优势基础上,提出了将信息进行匹配和调和相融合的处理方式。对实际的决策层信息融合目标识别问题,该算法具有弹性去除信息间相关性以及合理处理矛盾信息的能力。同时,MART神经网络模型通过自适应地调整网络参数,对信度的增长有较好的控制能力。 相似文献
7.
8.
Devore Michael D. O'Sullivan Joseph A. 《Multidimensional Systems and Signal Processing》2003,14(1-3):139-159
We present an approach to automatic target recognition (ATR) from synthetic aperture radar (SAR) imagery which combines advantages of both model-based and template-based approaches. Prior observations are used to estimate the statistical properties of reflectance over regions in the training scene. These target-centered statistical models can then be used to estimate the statistical properties of sensor output for arbitrary pose. Two-sided hypothesis tests which are maximally powerful at the most likely alternative are developed in a information-theoretic framework to address target model segmentation and confuser rejection. Segmentation of target from clutter is performed in the target-centered coordinate system using all prior observations to produce a consistent segmentation over all poses. We present performance and computation complexity results as a function of segmentation threshold, confuser-rejection threshold, and operating conditions for publicly available SAR data. 相似文献
9.
10.
基于多类特征融合的步态识别算法 总被引:2,自引:0,他引:2
步态识别作为一种新兴的生物认证技术,是指通过人的走路姿势来识别人的身份。由于步态受到很多因素的制约,基于单个特征识别率很低,而且不同的特征其特征类型和度量尺度不同。本文提出一种将人体轮廓特征、肢体角度特征、反射对称特征相融合的方法,得到一个联合特征矢量,并采用最近邻模糊分类器进行识别。实验结果表明,本文算法可以解决不同类别的特征融合问题,具有较好的识别性能。 相似文献
11.
在空间目标识别中有电磁、光度、红外、无线电等多种传感器信息可以利用,不同类型的传感器获得的信息在时域、频域和特征域上的冗余和互补,可以提高识别的正确率,扩展识别的时间和空间覆盖范围,而对于复杂的空间目标识别问题,不可避免的要用到多种分类方法,这些分类器的自动优化组合是融合识别获得良好性能的关键.本文分析了空间目标分布式传感器信息处理系统中的单传感器优缺点,并设计了基于多级增强融合结构的空间目标融合识别方案,给出了其系统级融合的软件设计.该方案能根据不同传感器和分类器的性质自动的对其进行优化组合,充分开发多分类器系统的潜力、提高异质多传感器融合识别的效率和稳健性. 相似文献
12.
13.
基于多特征融合的运动对象识别算法 总被引:1,自引:0,他引:1
为提高视频检索的准确率,提出了一种基于多特征融合的视频运动对象识别算法。该算法首先使用基于背景帧构造及关键帧截取的方法提取视频中的运动对象的区域;然后分别提取运动对象的局部特征SURF描述子和全局特征如颜色直方图、边缘直方图等,并使两者融合为统一的特征向量;最后使用支持向量机对特征进行学习和识别,用以识别视频对象。实验证明该算法有效地提高了视频中运动对象识别的准确率。 相似文献
14.
15.
16.
陈倩 《信息安全与通信保密》1994,(4)
日本NTT实验室开发的用于ISDN视听业务的加密板,符合国际电信联盟(ITU—T)的H.233建设。本文首先概述H.233建议,然后描述视听业务加密板的建立过程、结构及FEAL加密处理器的特点。 相似文献
17.
We have developed a real-time system to transform anaudio signal into several specialized representations of sound.The system uses analog circuit models of biological auditionto compute these representations. We report on a speech recognizerthat uses this system for feature extraction, and we evaluatethe performance of this speech recognition system on a speaker-independent13-word recognition task. 相似文献
18.
19.
从实用角度出发分析了相关滤波法、目标识别法和D-S证据理论等方法在异类传感器信息融合的不同层次的使用,综合以上方法实现对雷达侦察辐射源、通信辐射源、激光辐射源和雷达航迹的信息融合,最终合成平台信息,用于完成平台识别. 相似文献