近紅外人臉圖像的分類研究

1.24萬字
我自己的畢業(yè)論文，原創(chuàng)的，已經(jīng)通過校內(nèi)系統(tǒng)檢測(cè)，僅在本站獨(dú)家出售，重復(fù)率低，大家放心下載使用

摘要 近年來人們?cè)絹碓蕉嗟年P(guān)注生物特征識(shí)別，而在這些生物特征識(shí)別方法中，人臉識(shí)別具有方便，經(jīng)濟(jì)而準(zhǔn)確的特點(diǎn)，可以廣泛應(yīng)用于高安全性部門的警戒、入口控制、人機(jī)交互、計(jì)算機(jī)保密、公共安全等方面。在人臉識(shí)別中，人臉外觀會(huì)受到光照、姿態(tài)、表情變化的影響，所以人臉識(shí)別系統(tǒng)要適應(yīng)各種環(huán)境，此外化妝、照片欺詐也是人臉識(shí)別亟待解決的問題。本文主要研究基于二維線性判別分析的人臉識(shí)別。
本文首先介紹人臉識(shí)別的研究意義、識(shí)別方法、研究中的問題和難點(diǎn)以及人臉識(shí)別的現(xiàn)狀和發(fā)展前景。研究以近紅外圖像為基礎(chǔ)的人臉識(shí)別方法，并且也了解近紅外圖像人臉識(shí)別的特點(diǎn)和價(jià)值，它的缺點(diǎn)和可提高識(shí)別性能的研究方向也是要了解的。根據(jù)以上內(nèi)容，本文著重研究二維線性判別分析（2DLDA）方法。2DLDA算法識(shí)別率高、處理速度快，可以對(duì)人臉圖像數(shù)據(jù)降維，將人臉圖像測(cè)試樣本變換為二維矩陣并進(jìn)行列/行方向的二維線性判別分析。這是2DLDA算法的優(yōu)勢(shì)。
在Matlab上對(duì)基于二維線性判別分析的人臉識(shí)別算法進(jìn)行編程實(shí)現(xiàn)。
關(guān)鍵詞 近紅外 人臉識(shí)別 二維線性判別


Classification of near infrared face image

Abstract In recent years, more and more people were concerned about biometric identification, and in these biometric identification methods, face recognition with the characteristic of convenient, economical and accurate, can be widely used in high security alert authorities, access control, human interactive, computer secrecy, public safety and other aspects. In face recognition,light, gesture, facial expression affected face appearance, so the face recognition system have to adapt to the environment. In addition to makeup, photo recognition fraud were also serious problems. This paper studied the research of face recognition based on two-dimensional linear discriminant analysis.
This paper introduced the significance of face recognition, identification methods, research problems and difficulties as well as research status of face recognition and development prospects. The paper discussed the characteristics and values of face recognition based on the near-infrared image ​, comprehended the face recognition methods based on near-infrared image, meanwhile discussed the shortcomings of face recognition research based on the near-infrared image and improve recognition performance. On this basis, the paper focused on the two-dimensional linear discriminant analysis (TDLDA) method. 2DLDA algorithm had higher recognition rate, fast processing speed, can reduce the dimensionality of face image data, the face image test samples were converted into a two-dimensional matrix and thus made two-dimensional linear discriminant analysis of row/column direction. This is the advantages of 2DLDA algorithm.
A program about the algorithm of face recognition based on 2DLDA is written on MATLAB.
Key words Near-infrared facial recognition 2DLDA

目錄
第一章 緒論…………………………………………………………………………........1
1.1 研究人臉識(shí)別的意義……………………………………………..............................1
1.2 人臉識(shí)別的研究?jī)?nèi)容……………………………………………..............................1
1.3 人臉識(shí)別研究現(xiàn)狀及難點(diǎn)………………………………………..............................1
1.4 人臉識(shí)別系統(tǒng)……………………………………………………..............................2
1.5 人臉識(shí)別的發(fā)展趨勢(shì)及應(yīng)用領(lǐng)域………………………………..............................2
1.6 本文主要工作及結(jié)構(gòu)安排……………………………………………………...….3
第二章 近紅外人臉識(shí)別及二維線性判別分析………..………............................ 4
2.1 近紅外人臉識(shí)別……………………………………………………………...….....4
2.1.1 近紅外人臉識(shí)別的意義……………………………………………………...……4
2.1.2 近紅外圖像…………………………………………………………………...……4
2.2 人臉圖片庫(kù)……………………………………………………………………...….5
2.3 線性判別分析…………………………................ ………………...........................6
2.3.1 LDA與2DLDA………………………………………………………………........6
2.3.2 線性判別分析的缺點(diǎn)與改進(jìn)………………………………………………….......6
第三章 基于二維線性判別分析的人臉識(shí)別………………………………….…...8
3.1 線性判別分析的原理…………………………………………………………..........8
3.2 線性判別法……………………………………………………………………...….9
3.3 2DPCA原理……………………………………………………………...………….10
3.4 2DLDA原理……………………………………………………………...………….10
3.5 2DLDA在人臉識(shí)別中的應(yīng)用…………………………………….…………...……12
3.5.1 特征提取……………………………………………………………………….....12
3.5.2 分類…………………………………………………………………………..…...12
3.6 算法評(píng)價(jià)標(biāo)準(zhǔn)…………………………………………………………………….....13
第四章 MATLAB編程實(shí)現(xiàn)及運(yùn)行結(jié)果分析……………………………………... 14
4.1 MATLAB相關(guān)函數(shù)簡(jiǎn)介………..