跳到主要內容

臺灣博碩士論文加值系統

(3.235.227.117) 您好!臺灣時間:2021/08/01 23:27
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:葉又誠
研究生(外文):Yow-Cherng,Yeh
論文名稱:應用隱馬可夫模型及基因演算法於人臉辨識之研究
論文名稱(外文):A Face Recognition Scheme using Hidden Markov Models and Genetic Algorithm
指導教授:李建德李建德引用關係
指導教授(外文):Jiann-Der,Lee
學位類別:碩士
校院名稱:長庚大學
系所名稱:電機工程研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2005
畢業學年度:93
語文別:中文
中文關鍵詞:隱馬可夫模型基因演算法
外文關鍵詞:Hidden Markov ModelsGenetic Algorithm
相關次數:
  • 被引用被引用:1
  • 點閱點閱:241
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:1
在人臉辨識的研究領域裡,傳統上有使用如類神經網路、幾何特徵法、以樣版為基礎法、相關係數法、eigenface等,這些方法各有其優缺點,其中有一種方法使用隱馬可夫模型(Hidden Markov Models),而其特徵擷取,一般採用2D-DCT及2D-DWT二種方式,輸入的資料為灰階影像,使用這種方法的辨識率有不錯的結果,但為了更進一步提升辨識率,本研究提出一種改良方式去取代傳統灰階影像產生方式,將輸入資料改為彩色影像並且在傳統隱馬可夫模型辨識架構加上基因演算法,以基因演算法搜尋影像最佳色彩比例轉換,因而可以提升辨識率,因為基因演算法可以加強人臉區域特徵值。為了驗證這種方法,我們的實驗首先採用傳統HMM,特徵擷取方法採用2D-DCT及2D-DWT,輸入資料採用ORL、CVL及本實驗室人臉影像資料庫於實驗,選定一組最佳的2D-DWT,最後將這一組方法加上基因演算法做驗證實驗,經由這些實驗證明,基因演算法加傳統HMM可以提升原來的辨識率,從實驗數據中得知,本實驗室人臉資料庫的辨識率比起使用傳統HMM方法,最高可提升12%,平均提升量大約為4.67%,平均辨識率為94.4%,CVL人臉資料庫的辨識率最高可提升10%,平均提升量為10%,平均辨識率為80%,並且由實驗結果得知基因演算法把原來影像平滑部分抑制,邊緣部分強調出來,所以可以提升辨識率;因此把基因演算法與其他辨識方法作結合,這構成一個有趣值得深入研究的方向。
Among face recognition researches, most methods can be grouped as the approaches: using neural networks, using geometry features, using templates, using correlation coefficients, using Hidden Markov Models (HMM), and using eigen-face. Every method has its advantages and disadvantages as mentioned in their original papers, and the most popular method is the approach using HMM with the extracted features in 2-D DCT and DWT domain in the last decades. Many researches using HMM have been successfully shown promised recognition results with different features such as a new block scan path or using different filters. However, they only considered the traditional gray images like Y component of YUV format as their input images. To improve their results further, a different approach is proposed in this research that Genetic Algorithm (GA) is cooperated with HMM to search the best combined ratio of RGB images and generate a new gray input image with emphasized local face features. In the experiment results using ORL, CVL and my co-workers’ facial images with different captured angels and emotions, the best improvement rate of the proposed method using GA and HMM is 12% and the average rate is 4.67% on my co-workers’ facial data compared with the traditional HMM methods. On the CVL facial data, the best performance of improvement rate is 10%. Genetic Algorithm improve the recognition rates because the new gray input images with enhanced local features such as stronger edge information generated when compared with the traditional gray input images. The concept using GA to enhance the input facial images indicates a new further research direction to improve the conventional HMM methods.
第一章 緒論....................................................1
1.1 前言................................................ 1
1.2 研究動機............................................ 1
1.3 論文架構............................................ 2
第二章 文獻回顧................................................3
2.1 隱馬可夫模型的基本觀念.............................. 3
2.2 隱馬可夫模型的數學原理.............................. 4
第三章 隱馬可夫模型應用在影像辨識.............................13
3.1 建構人臉影像的HMM................................13
3.2 特徵擷取........................................... 13
3.2.1 信號取樣 ......................................14
3.2.2 信號轉換.......................................15
3.2.3 取出特徵值.....................................21
3.3 編碼............................................... 21
3.4 模型的建立與辨識................................... 22
第四章 基因演算法.............................................25
4.1 簡介............................................... 25
4.2 基因演算法流程..................................... 25
4.3 基因演算法說明..................................... 26
第五章 實驗結果...............................................32
5.1 不同狀態數辨識率結果............................... 33
5.2不同重疊量辨識率結果................................34
5.3 HMM+2D-DWT & 2D-DCT ............................35
5.4 HMM+Lifting........................................36
5.5 特徵維度(Feature dimension) 的影響....................36
5.6 遮蔽測試........................................... 37
5.7 以基因演算法求最佳色彩比例......................... 38
5.8 應用程式介面....................................... 44
第六章 結論與未來研究方向 ....................................48
參考文獻......................................................50
[1] S.Naderi, M.S. Moin, N.M. Char kari,” 1D-HMM for face verification: model optimization using improved algorithm and intelligent selection of training images”, Proceedinds of International Conference of Pattern Recognition (ICPR), vol. 3, pp. 330-333, 2004.
[2] Lawrence R. Rabiner, ”A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition”, Proceedings of the IEEE, vol. 77 , no. 2, pp. 257-286 ,1989
[3] 陳同孝, 張真誠, 黃國峰, “數位影像處理技術”, 旗標出版社, 2004
[4] Konstantin Markov , Satoshi Nakamura, “A Hybrid HMM/BN Acoustic Model for Automatic Speech Recognition,” IEICE Trans. Inf. & Syst., vol. E86-D, no. 3, pp. 438–445, 2003
[5] Ara V. Nefian and Monson H. Hayes III,” Hidden Markov Models For Face Recognition”, Acoustics, Speech, and Signal Processing, ICASSP, vol. 5, pp. 2721-2724, 1998.
[6] V. V. Kohir, U.B. Desai, ”Face recognition using a DCT-HMM approach”, Applications of ComputerVision , WACV, pp. 226–231 , 1998.
[7] M. Bicego, U. Castellani, V. Murino,” Using Hidden Markov Models and Wavelets for Face Recognition”, Image Analysis and Processing, pp.52-56, 2003.
[8] W. Sweldens, “The lifting scheme: a custom-designed construction of biorthogonal wavelets”, Applied and Computational Harmonic Analysis, vol. 3, no. 2, pp. 186-200, 1996
[9] R. He, P.A. Narayana,”Global optimization of mutual information: application to three-dimensional retrospective registration of magnetic resonance images”, Computerized Medical Imaging and Graphics, vol. 26 , pp. 277-292 , 2002.
[10] 張啟仲, “利用基因演算法改善自動網格切割”, 碩士論文, 中山大學電機工程研究所, 2004
[11] ORL Face Database: http://www.uk.research.att.com/pub/data/att_faces.zip
[12] CVL Face Database: http://www.lrv.fri.uni-lj.si/facedb.html
[13] Franc Solina, Peter Peer, Borut Batagelj, Samo Juvan, Jure Kovac, “Color-based face detection in the "15 seconds of fame" art installation”, Proceedings of Mirage 2003, INRIA Rocquencourt, France, pp. 38-47, 2003
[14] Ara V. Nefian and Monson H. Hayes III , ”Face detection an recognition using hidden Markov models” , Image Processing, ICIP, vol. 1, pp. 141-145, 1998.
[15] Hung-Son Le, Haibo Li, ”Face identification system using single hidden Markov model and single sample image per person”, Neural Networks, vol. 1, pp. 25-29, 2004
[16] V.V. Kohir, U.B. Desai,” A transform domain face recognition approach”, Proceedinds of IEEE TENCON 1999, vol. 1, pp. 104-107, 1999
[17] S. Eickeler, S. Muller, G. Rigoll, “Recognition of JPEG compressed face images based on statistical methods”, Image and Vision Computing , vol. 18, pp. 279-287, 2000.
[18] F.S. Samaria , A.C. Harter , ” Parameterisation of a stochastic model for human face identification”, Proceedinds of Applications of Computer Vision, pp. 138-142, 1994
[19] H. Othman, T. Aboulnasr,” A separable low complexity 2D HMM with application to face recognition”, Pattern Analysis and Machine Intelligence, vol. 25, pp. 1229-1238, 2003
[20] Valery A. Petrushin , ”Hidden Markov Models: Fundamentals and Applications”, Online Symposium for Electronics Engineers 2000 (http://www.techonline.com/osee/) , 2001
[21] Eric Fosler-Lussier, “Markov Models and Hidden Markov Models: A Brief Tutorial”, International Computer Science Insitute Technical Reports, tr-98-041, December 1998
[22] Horst Bunke, Terry Caelli, ”Hidden Markov Models Application in Computer Vision”, World Scientific, 2001
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top