Basic Search / Detailed Display

Author: 楊智州
Chin-Chou Yang
Thesis Title: 結合卷積神經網路與K-近鄰演算法之行動裝置人臉辨識系統
A Face Recognition System Based on Mobile Device using Convolutional Neural Network and K-Nearest Neighbors
Advisor: 洪西進
Shi-Jinn Horng
Committee: 郭重顯
李正吉
吳怡樂
林韋宏
Degree: 碩士
Master
Department: 電資學院 - 資訊工程系
Department of Computer Science and Information Engineering
Thesis Publication Year: 2018
Graduation Academic Year: 106
Language: 中文
Pages: 55
Keywords (in Chinese): 生物辨識人臉辨識影像處理卷積神經網路K-近鄰演算法行動裝置
Keywords (in other languages): Biometric, Face recognition, Image processing, Convolutional Neural Network, K Nearest Neighbor, Mobile device
Reference times: Clicks: 344Downloads: 0
Share:
School Collection Retrieve National Library Collection Retrieve Error Report

人體的生物特徵可區分為生理特徵和行為特徵,生理特徵如:臉部、指紋、靜脈、掌紋、骨骼、虹膜等,屬於靜態的特徵,行為特徵如:簽名、聲紋、心跳、步態等,則屬於動態的特徵,這些生物特徵都具有個體的唯一性、穩定性,因此,可利用這些特徵作為判別個人身分的依據,發展出各式的生物辨識技術。同時,隨著資訊科技與網際網路的發展,行動裝置的已經非常普及,且應用的範圍廣泛,個人資訊安全更顯得重要,使得生物辨識技術結合行動裝置在安全驗證上受到重視。
這些生物辨識技術中,人臉辨識最大的特點就是自然性和非接觸性。自然性指的是和人類進行個體辨識時所利用的生物特徵相同,例如:人臉辨識是通過觀察比較人臉對其他個體進行區分和確認身分。除此之外,使用人臉作為生物特徵辨識擁有取得容易,不用接觸即可完成的優點,故成為主流的生物辨識方法之一。
本論文在人臉偵測部分是使用哈爾特徵篩選出人臉的影像,再以尺度正規化、旋轉正規化、亮度正規化對影像的進行前處理,並使用卷積神經網路(Convolutional Neural Network, CNN)作為特徵擷取的方法,最後使用K-近鄰演算法(K Nearest Neighbor, KNN) 作為辨識比對的方法,並將此系統架構建立在行動裝置上使用。


Biological features of the human can be divided into physiological features and behavioral features, physiological features such as face, fingerprint, vein, palmprint, bone, iris, etc., belong to the static features, behavioral features such as signature, voiceprint, heartbeat, gait, etc., belong to the dynamic features, these biological features has individual uniqueness and stability, therefore, we can use these features as determining personal identity, develop each type of biometric technology. Meanwhile, with information technology and internet development, mobile device has been very popular, and the range of applications is extensive, personal information security has become more important, so that the biometric technology combined with the mobile device in the security verification attention on.
In these biometric techniques, face recognition features are natural and non-contact. Naturalness refers to a human individual identification by using biological characteristics. For example, face recognition is comparing faces between different individuals to distinguish and confirm the identity. In addition, using faces as a biometric identification is easy and without contact, so it becomes the mainstream of biometric methods.
We propose a face recognition system based on mobile device using haar-like feature to detect human face. And preprocessing the image with the scale normalization, rotation normalization and illumination normalization. And use Convolutional Neural Network as the method to extract feature. Then use K Nearest Neighbor as a method of recognition.

摘要 Abstract 致謝 目錄 圖目錄 表目錄 第一章 緒論 1.1 研究動機與目的 1.2 相關研究回顧 1.3 論文章節介紹 第二章 系統架構與流程 2.1 系統架構 2.2 系統流程 2.2.1 註冊流程 2.2.2 辨識流程 第三章 人臉偵測與影像前處理 3.1 人臉偵測(Face Detection) 3.1.1 哈爾特徵(Haar-Like Features) 3.1.2 積分影像(Integral Image) 3.1.3 AdaBoost演算法(AdaBoost Algorithm) 3.1.4 級聯分類器(Cascade Classifiers) 3.2 人臉影像前處理 (Face Image Preprocessing) 3.2.1 尺度正規化(Scale Normalization) 3.2.2 旋轉正規化(Rotation Normalization) 3.2.3 亮度正規化 (Illumination Normalization) 第四章 影像特徵擷取與比對 4.1 特徵擷取(Feature Extraction) 4.1.1 卷積神經網路(Convolutional Neural Network, CNN) 4.1.2 改良後所使用之卷積神經網路 4.2 影像特徵描述(Image Features Representation) 4.3 特徵比對(Feature Matching) 第五章 實驗結果 5.1 系統開發環境 5.2 系統執行畫面 5.3 人臉資料庫(Face Database) 5.4 實驗結果(Experimental Results) 第六章 結論 參考文獻

[1] Consumer Barometer with Google, https://www.consumerbarometer.com/en/trending/?countryCode=TW&category=TRN-NOFILTER-ALL
[2] Amjad, A., Griffiths, dA., Patwary, M.N.: “Multiple face detection algorithm using colour skin modelling”, IET Image Processing, 2012
[3] Sagheer, A., Aly, S.: “An Effective Face Detection Algorithm Based on Skin Color Information”, SITIS, 2012
[4] Kang-Seo Park, Rae-Hong Park, Young-Gon Kim: “Face detection using the 3×3 block rank patterns of gradient magnitude images and a geometrical face model”, ICCE, 2011
[5] Wei Chen, Tongfeng Sun, Xiaodong Yang, Li Wang: ”Face detection based on half face-template”, ICEMI , 2009
[6] Wang Jizeng, Yang Hongmei: “Face Detection Based on Template Matching and 2DPCA Algorithm”, CISP, 2008
[7] Amaro E.G., Nuno-Maganda M.A., Morales-Sandoval M.: “Evaluation of machine learning techniques for face detection and recognition”, CONIELECOMP, 2012
[8] Rainer Lienhart, Jochen Maydt: “An Extended Set of Haar-like Features for Rapid Object Detection”, ICIP, 2002
[9] Mita, T., Kaneko, T., Hori, O.: “Joint Haar-like features for face detection”, ICCV, 2005
[10] Viola, P., Jones, M.: “Robust Real-Time Face Detection”, IJCV, 2004,
[11] Viola, P., Jones, M.: “Rapid object detection using a boosted cascade of simple features”, CVPR, 2001
[12] Chunhua Shen, Paisitkriangkrai, S., Zhang, Jian: “Efficiently Learning a Detection Cascade With Sparse Eigenvectors”, IEEE Transactions on Image Processing, 2010
[13] Ning Jiang, Wenxin Yu, Shaopeng Tang, Goto, S.: “A cascade detector for rapid face detection”, CSPA, 2011
[14] Rong Xiao, Huaiyi Zhu, He Sun, Xiaoou Tang: “Dynamic Cascades for Face Detection”, ICCV, 2007
[15] Lee, P.-H., Wu, S.-W., Yi-Ping Hung: “Illumination Compensation Using Oriented Local Histogram Equalization and its Application to Face Recognition”, IEEE Transactions on Image Processing, 2012
[16] Sharif, M., Mohsin, S., Jamal, M.J., Raza, M.: “Illumination normalization preprocessing for face recognition”, ESIAT, 2010
[17] Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton: “ImageNet Classification with Deep Convolutional Neural Networks”, NIPS ,2012
[18] D.W. Ruck, S.K. Rogers, M.Kabrisky, M.E. Oxley, B.W. Sutert: “The multilayer perceptron as an approximation to a Bayes optimal discriminant function”, IEEE Transactions on Neural Networks, 1990
[19] Wikipedia:https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#GeForce_10_seriesList of Nvidia graphics processing units
[20] GPU GFLOPS : https://gflops.surge.sh/
[21] Forest Agostinelli, Matthew Hoffman, Peter Sadowski, Pierre Baldi:“Learning Activation Functions to Improve Deep Neural Networks”, ICLR, 2015
[22] Bing Xu, Naiyan Wang, Tianqi Chen, Mu Li:“Empirical Evaluation of Rectified Activations in Convolutional Network”, ICML, 2015
[23] Andrew L. Maas, Awni Y. Hannum, Andrew Y. Ng:“Rectifier Nonlinearities Improve Neural Network Acoustic Models”, ICML, 2013
[24] Altman, N. S.:“An introduction to kernel and nearest-neighbor nonparametric regression”, The American Statistician, 1992
[25] Deza Elena, Deza Michel Marie: “Encyclopedia of Distances”, Springer , 2009
[26] “Labeled Faces in the Wild database”, http://vis-www.cs.umass.edu/lfw/
[27] Sun Y, Wang X, Tang X: “Deep learning face representation from predicting 10,000 classes”, CVPR , 2014
[28] Sun Y, Wang X, Tang X: “Deep Learning Face Representation by Joint Identification-Verification”, CVPR, 2015
[29] Chaochao Lu, Xiaoou Tang: “Surpassing Human-Level Face Verification Performance on LFW with GaussianFace”, AAAI, 2014
[30] Yaniv Taigman, Ming Yang, Marc'Aurelio Ranzato, Lior Wolf : “DeepFace: Closing the Gap to Human-Level Performance in Face Verification”, CVPR , 2014
[31] Xudong Cao, David Wipf, Fang Wen, Genquan Duan, Jian Sun: “A Practical Transfer Learning Algorithm for Face Verification”, ICCV, 2013
[32] Dong Chen, Xudong Cao, Fang Wen, Jian Sun: “Blessing of Dimensionality: High-Dimensional Feature and Its Efficient Compression for Face Verification”, CVPR , 2013

無法下載圖示 Full text public date 2023/08/25 (Intranet public)
Full text public date This full text is not authorized to be published. (Internet public)
Full text public date This full text is not authorized to be published. (National library)
QR CODE