| 查看: 1041 | 回复: 3 | ||
| 当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖 | ||
likaiaiswt新虫 (小有名气)
|
[求助]
人脸识别翻译！
|
|
|
PCA: Principal Components Analysis (PCA) PCA, commonly referred to as the use of eigenfaces, is the technique pioneered by Kirby and Sirivich in 1988. With PCA, the probe and gallery images must be the same size and must first be normalized to line up the eyes and mouth of the subjects within the images. The PCA approach is then used to reduce the dimension of the data by means of data compression basics2 and reveals the most effective low dimensional structure of facial patterns. This reduction in dimensions removes information that is not useful4 and precisely decomposes the face structure in toorthogonal (uncorrelated) components known as eigenfaces. Each face image may be represented as a weighted sum (feature vector) of the eigenfaces, which are stored in a 1D array. A probe image is compared against a gallery image by measuring the distance between their respective feature vectors. The PCA approach typically requires the full frontal face to be presented each time; otherwise the image results in poor performance.4 The primary advantage of this technique is that it can reduce the data needed to identify the individual to 1/1000th of the data presented. 5This Document Last Updated: 7 August 2006 Page 2 of 10 Face Recognition Figure 1: Standard Eigenfaces: Feature vectors are derived using eigenfaces.6 LDA: Linear Discriminant Analysis LDA is a statistical approach for classifying samples of unknown classes based on training samples with known classes.4 (Figure 2)This technique aims to maximize between-class (i.e., across users) variance and minimize within-class (i.e., within user)variance. In Figure 2 where each block represents a class, there are large variances between classes, but little variance within classes. When dealing with high dimensional face data, this technique faces the small sample size problem that arises where there are a small number of available training samples compared to the dimensionality of the sample space.7 Figure 2: Example of Six Classes Using LDA 8This Document Last Updated: 7 August 2006 Page 3 of 10Face Recognition EBGM: Elastic Bunch Graph Matching EBGM relies on the concept that real face images have many nonlinearcharacteristics that are not addressed by the linear analysismethods discussed earlier, such as variations in illumination(outdoor lighting vs. indoor fluorescents), pose (standing straightvs. leaning over) and expression (smile vs. frown). A Gaborwavelet transform creates a dynamic link architecture thatprojects the face onto an elastic grid.4 The Gabor jet is a node onthe elastic grid, notated by circles on the image below, whichdescribes the image behavior around a given pixel. It is the resultof a convolution of the image with a Gabor filter, which is used todetect shapes and to extract features using image processing. [Aconvolution expresses the amount of overlap from functions,blending the functions together.] Recognition is based on thesimilarity of the Gabor filter response at each Gabor node.4 Thisbiologically-based method using Gabor filters is a processexecuted in the visual cortex of higher mammals. The difficultywith this method is the requirement of accurate landmarklocalization, which can sometimes be achieved by combining PCAand LDA methods.4 Figure 4: Elastic Bunch Map Graphing.9 |
回复此楼» 猜你喜欢
考研调剂
已经有7人回复
-
0856调剂
已经有4人回复
-
材料277求调剂
已经有5人回复
-
一志愿北京化工大学材料与化工（085600）296求调剂
已经有6人回复
-
资源与环境 调剂申请(333分)
已经有8人回复
-
334分 一志愿武理 材料求调剂
已经有3人回复
-
284求调剂
已经有5人回复
-
0703化学/290求调剂/本科经历丰富/工科也可
已经有6人回复
-
北科281学硕材料求调剂
已经有18人回复
-
275求调剂
已经有9人回复
fjtony163
版主 (文坛精英)
米米
- 翻译EPI: 386
- 应助: 361 (硕士)
- 贵宾: 6.876
- 金币: 70610.6
- 散金: 36050
- 红花: 222
- 沙发: 616
- 帖子: 21912
- 在线: 3793.7小时
- 虫号: 1397876
- 注册: 2011-09-11
- 专业: 遗传学研究新技术与方法
- 管辖: 论文翻译
4楼2012-12-26 19:01:36
sklyer_123
银虫 (小有名气)
- 翻译EPI: 2
- 应助: 4 (幼儿园)
- 金币: 277.4
- 红花: 1
- 帖子: 241
- 在线: 12.5小时
- 虫号: 2198228
- 注册: 2012-12-20
- 性别: MM
- 专业: 会计与审计
【答案】应助回帖
爱与雨下: 请解释一下 2012-12-26 21:46:02
|
主成分分析：主成分分析（常设仲裁法院） 主成分分析，通常称为使用的特征，是首创的技术和sirivich柯比在1988。主成分分析，探头和画廊的图片必须是相同的大小，必须首先正常化线的眼睛和嘴的科目内的图像。主成分分析法是用来减少维度的数据通过数据压缩basics2揭示最有效的低维结构的面部形态。这减少了尺寸，消除信息不useful4和精确分解结构在toorthogonal（无关）部件被称为特征。每个人脸图像可以表示为一个加权（特征向量）的特征，这是存储在一个一维数组。探针图像比对一个画廊图像之间的距离测量它们各自的特征向量。主成分分析法通常需要全脸是每次；否则图像导致性能较差。4主要利用这一技术，它可以减少所需的数据确定个人1 /第一千的数据。 这文件最后更新：7威严的2006页2 10 图1：标准特征脸的人脸识别：特征向量推导使用6的特征。 分析：线性判别分析 是一种统计方法进行分类未知样品类别根据训练样本与已知类别。4（图2）这种技术的目的是最大限度地类（即，在用户）方差最小化类内（即，在用户）方差。在图2中，每一块代表一个阶级，有很大的差异，阶级之间，但很少差额内部类。在处理高维数据，这种技术所面临的小样本问题出现在有少量训练样本的比较维度的样本空间。7 图2：例如六类激光8文件最后更新：7威严的2006页3 10face识别 弹性图像匹配算法：弹性束图匹配 弹性图像匹配算法依赖的概念，真正的人脸图像有许多nonlinearcharacteristics，没有处理的线性分析方法讨论，如光照的变化（户外照明和室内荧光灯），姿势（站straightvs。俯身）和表达（笑比皱眉）。一个gaborwavelet变换创建动态链接结构thatprojects面临到一个网格。4该射流是一个节点在弹性网格，标记圆圈下面的图像，whichdescribes图像的行为在一个给定的像素。这后者是卷积的图像提供了一个滤波器，这是用来检测的形状和特征提取的图像处理。[ aconvolution表示数量的重叠的功能，融合功能。]识别是基于thesimilarity的伽柏滤波器响应每个伽柏节点。4 thisbiologically-based方法使用滤波器是一个processexecuted在视觉皮层较高哺乳动物。该difficultywith是这种方法的要求，准确landmarklocalization，有时可通过结合4 pcaand测速方法。 图4：弹性束图9图表。  |
2楼2012-12-25 16:44:44
likaiaiswt
新虫 (小有名气)
- 应助: 0 (幼儿园)
- 金币: 495
- 帖子: 271
- 在线: 137.5小时
- 虫号: 2104068
- 注册: 2012-11-03
- 性别: GG
- 专业: 历史社会学
3楼2012-12-25 22:03:26
