本文選題：Hadoop　切入點(diǎn)：多姿態(tài)人臉識(shí)別　出處：《吉林大學(xué)》2017年碩士論文

【摘要】：隨著社會(huì)的不斷發(fā)展,科技水平的不斷進(jìn)步,人臉識(shí)別技術(shù)由于具有結(jié)果直觀、隱藏性好、操作方便的優(yōu)越性,被廣泛應(yīng)用于信息安全、金融安全、反恐環(huán)境、刑偵調(diào)查等領(lǐng)域。但是在非約束環(huán)境下,攝像頭往往采集不到非常合適的人臉,比如面部遮擋,復(fù)雜環(huán)境干擾,表情、姿態(tài)的變化,都會(huì)在一定程度上導(dǎo)致人臉識(shí)別效果下降。另外,伴隨著不斷擴(kuò)大規(guī)模的數(shù)據(jù)生成點(diǎn),圖像數(shù)據(jù)年都以20%的增長(zhǎng)率快速增加,如何在龐大的數(shù)據(jù)中快速的查找到目標(biāo)人臉圖像,這就需要將云計(jì)算服務(wù)應(yīng)用在傳統(tǒng)的識(shí)別技術(shù)中。所以,為了解決以上出現(xiàn)的問題,本文以Hadoop結(jié)構(gòu)為基礎(chǔ),構(gòu)建了云計(jì)算平臺(tái),并在Hadoop云計(jì)算平臺(tái)下對(duì)非約束環(huán)境中多姿態(tài)人臉識(shí)別展開了研究。本文主要從以下幾個(gè)方面展開研究工作:1、研究了目前流行的Hadoop云平臺(tái)結(jié)構(gòu),以VMware workstation作為虛擬機(jī),創(chuàng)建了3個(gè)ubuntu系統(tǒng),并將3個(gè)ubuntu系統(tǒng)組成Master/Slaver結(jié)構(gòu),同時(shí)配置好相關(guān)的文件、網(wǎng)絡(luò)和軟件,最后構(gòu)建Hadoop完全分布式系統(tǒng)。2、研究了傳統(tǒng)的神經(jīng)網(wǎng)絡(luò)的組成與特點(diǎn),然后進(jìn)一步依據(jù)神經(jīng)網(wǎng)絡(luò)特點(diǎn)分析卷積神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu),并且主要對(duì)卷積神經(jīng)網(wǎng)絡(luò)的卷積層、池化層和激活層進(jìn)行了進(jìn)一步的研究,最后簡(jiǎn)要總結(jié)了卷積神經(jīng)網(wǎng)絡(luò)的特點(diǎn)。3、通過對(duì)當(dāng)前流行的Le Net-5卷積神經(jīng)網(wǎng)絡(luò)模型進(jìn)行性能試驗(yàn)與分析,研究卷積神經(jīng)網(wǎng)絡(luò)的卷積核尺寸與數(shù)目、正則化方式、激活函數(shù)、池化方式等參數(shù)對(duì)該模型相關(guān)性能的影響,從而能夠選取最合適的模型參數(shù)來構(gòu)建出最優(yōu)的卷積神經(jīng)網(wǎng)絡(luò)模型,并且在CAS-PEAl人臉庫上進(jìn)行了驗(yàn)證。4、利用本文搭建的Hadoop云平臺(tái)進(jìn)行圖片集的收集與整理,然后使用本文改進(jìn)的卷積神經(jīng)網(wǎng)絡(luò)對(duì)收集到的圖片集合進(jìn)行特征提取,并通過PCA算法將提取的特征向量進(jìn)行降維,最后識(shí)別階段采用余弦相似度度量算法進(jìn)行目標(biāo)人臉識(shí)別。本實(shí)驗(yàn)在CAS-PEAl人臉庫進(jìn)行,并對(duì)本文改進(jìn)的多姿態(tài)人臉識(shí)別算法在識(shí)別正確率與識(shí)別時(shí)間上進(jìn)行了有效的分析。
[Abstract]:With the development of society and the progress of science and technology, face recognition technology has been widely used in information security, financial security, anti-terrorism environment because of its advantages of intuitive results, good concealment and convenient operation. But in unconstrained environments, cameras often fail to capture very suitable human faces, such as facial occlusion, complex environmental disturbances, facial expressions, changes in posture, In addition, with the increasing scale of data generation point, the annual growth rate of image data increases rapidly by 20%, how to find the target face image quickly in the huge data. Therefore, in order to solve the above problems, this paper constructs cloud computing platform based on Hadoop structure. The multi-pose face recognition in unconstrained environment is studied on the Hadoop cloud computing platform. This paper mainly studies the structure of the popular Hadoop cloud platform in the following aspects: 1. VMware workstation is used as the virtual machine. Three ubuntu systems are created, and three ubuntu systems are formed into Master/Slaver structure. At the same time, the related files, network and software are configured. Finally, a fully distributed Hadoop system .2is constructed, and the composition and characteristics of traditional neural networks are studied. Then the structure of convolution neural network is analyzed according to the characteristics of neural network, and the convolution layer, pool layer and activation layer of convolutional neural network are studied further. Finally, the characteristics of convolution neural network are summarized briefly. Through the performance test and analysis of the current Le Net-5 convolution neural network model, the size and number of convolution cores, regularization mode, activation function of the convolutional neural network are studied. The influence of the parameters such as pool mode on the performance of the model can be used to select the most suitable model parameters to construct the optimal convolution neural network model. And on the CAS-PEAl face database validation. 4, using the Hadoop cloud platform built in this paper to collect and organize the picture set, and then use the improved convolution neural network to extract the features of the collected image set. The extracted feature vector is reduced by PCA algorithm, and the cosine similarity measure algorithm is used in the final recognition stage. This experiment is carried out in the CAS-PEAl face database. And the improved multi-pose face recognition algorithm is analyzed effectively in the recognition accuracy and recognition time.
【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 Guang-hui SONG;Xiao-gang JIN;Gen-lang CHEN;Yan NIE;;Two-level hierarchical feature learning for image classification[J];Frontiers of Information Technology & Electronic Engineering;2016年09期

2 勞定雄;;云計(jì)算在安防管理平臺(tái)中的應(yīng)用[J];中國(guó)公共安全;2013年14期

3 喬森;;基于云技術(shù)的數(shù)字化校園安全方案[J];電子制作;2013年08期

4 王伊卓;;暢想云計(jì)算環(huán)境下的云圖書館[J];河北科技圖苑;2013年01期

5 李翔;劉文兵;馬超;劉海艷;;基于云計(jì)算的空間數(shù)據(jù)處理技術(shù)[J];測(cè)繪與空間地理信息;2012年09期

6 郭麗;劉磊;季鐸;蔡?hào)|風(fēng);;基于奇異值分解的關(guān)系相似度計(jì)算方法[J];沈陽航空工業(yè)學(xué)院學(xué)報(bào);2009年03期

7 黃華,張建剛,許魯;藍(lán)鯨分布式文件系統(tǒng)的分布式分層資源管理模型[J];計(jì)算機(jī)研究與發(fā)展;2005年06期

8 張曉華,山世光,曹波,高文,周德龍,趙德斌;CAS-PEAL大規(guī)模中國(guó)人臉圖像數(shù)據(jù)庫及其基本評(píng)測(cè)介紹[J];計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào);2005年01期

9 陳粟,倪林;一種特征臉分析和小波變換相結(jié)合的人臉識(shí)別方法[J];計(jì)算機(jī)應(yīng)用;2004年10期

10 王雪光,郭艷兵,齊占慶;激活函數(shù)對(duì)BP網(wǎng)絡(luò)性能的影響及其仿真研究[J];自動(dòng)化技術(shù)與應(yīng)用;2002年04期

，

本文編號(hào)：1674415