【摘 要】
：
提出了一种基于特征融合的人脸识别方法.该方法首先对预处理后的人脸图像进行全局特征及局部分量的提取,分别采用离散余弦变换(DCT)提取包含图像大量信息的低频部分特征和奇异值分解(SVD)抽取图像的代数特征作为图像的全局特征,采用非负矩阵分解(NMF)提取图像的局部分量特征,然后将此两类特征以独立成份分析(ICA)进行融合,获取用于人脸识别的特征向量.在本文的实验中,我们将此特征向量应用于支持向量机(
论文部分内容阅读
提出了一种基于特征融合的人脸识别方法.该方法首先对预处理后的人脸图像进行全局特征及局部分量的提取,分别采用离散余弦变换(DCT)提取包含图像大量信息的低频部分特征和奇异值分解(SVD)抽取图像的代数特征作为图像的全局特征,采用非负矩阵分解(NMF)提取图像的局部分量特征,然后将此两类特征以独立成份分析(ICA)进行融合,获取用于人脸识别的特征向量.在本文的实验中,我们将此特征向量应用于支持向量机(SVM)进行分类训练及识别测试,并获得较好的结果.
其他文献
In order to realize the energy saving of the backlight in liquid crystal display (LCD) with approximate image quality,a human visual characteristics guided backlight luminance scaling method is propos
IASP 2010 aims to provide a high-level international forum for scientists, researchers and engineers to present the state-of-the-art advancements in the studies of Image Analysis and Signal Processing
This paper designs a 3 mm radiometer and validate with experiments based on the principle of passive millimeter wave(PMMW) imaging.The poor spatial resolution,which is limited by antenna size,should b
By using two-directional S-rough sets,the concepts of(f,f)-interference generation and(f,f)-inter-ference law generation of knowledge,F-interference generation and F-interference law generation of two
Moving object extraction and classification are important problems in automated video surveillance systems.A background model based on region segmentation is proposed.An adaptive single Gaussian backg
A simple and effective greedy algorithm for image approximation is proposed.Based on the matching pursuit approach,it is characterized by a reduced computational complexity benefiting from two major m
为了克服陶瓷产品信息发布的二维平面性、单一性问题,提出依托陶瓷企业现有电子商务网站。运用Web3D技术中的Cult3D和Virtools技术,构建三维在线互动形式的陶瓷产品虚拟展示平台,发布陶瓷产品信息,给客户以真实感。并介绍了其实现的关键技术,以陶瓷茶壶、花瓶为例进行三维互动展示。
An improved scheme for passive synthetic aperture imaging is proposed. With this new technique,the radiation signal from field of view (FOV) is received and its frequency is down-converted to intermed
提出了一种基于图像配准的自动目标识别算法,图像配准算法采用基于归一化互信息相似性判据,并采用模糊自适应粒子群优化算法作为搜索策略。在图像精确配准的基础上,通过图像间的相互转换,间接实现了目标的准确识别。仿真试验结果表明,该方法可以实现复杂背景下目标的准确识别。
针对六边形快速算法只考虑矢量水平方向的分布特点,提出了自适应六边形快速运动估计算法。文中用一个标量来统计水平和垂直运动矢量分布强弱特性,根据这个标量来选择调用六边形水平和垂直模板。综合考虑水平和垂直方向的运动矢量,降低计算量。实验表明,新算法在信噪比和搜索速度都有改进。