基于層次模型和融合決策的多姿態(tài)人臉識(shí)別技術(shù)
Multi-pose face recognition based on a hierarchical model and fusion decision
-
摘要: 該文提出了基于層次模型和融合決策的多姿態(tài)人臉識(shí)別技術(shù),它首先把各人臉按姿態(tài)分成幾個(gè)大類(lèi),并且對(duì)各大類(lèi)按人臉個(gè)體分成相應(yīng)子類(lèi),然后對(duì)各個(gè)大類(lèi)分別進(jìn)行基于特征臉的人臉識(shí)別,最后對(duì)各個(gè)姿態(tài)的人臉識(shí)別中間結(jié)果進(jìn)行融合決策得到真正的人臉識(shí)別結(jié)果,該算法同時(shí)也提供了其姿態(tài)識(shí)別結(jié)果,并且大大減小了耗時(shí),該文算法對(duì)ORL,UMIST,Stirling數(shù)據(jù)和一些自拍數(shù)據(jù)共1200幅人臉圖像進(jìn)行了識(shí)別測(cè)試實(shí)驗(yàn),其結(jié)果令人鼓舞。
-
關(guān)鍵詞:
- 人臉識(shí)別; 多姿態(tài); 特征臉; 多視圖; 層次模型; 融合決策
Abstract: A multi-pose face recognition algorithm based on a hierarchical model and fusion decision is proposed in the paper. First, face images are divided into several classes based on their poses and every class into many sub-classes based on their subjects. Then, traditional face recognition technology based on eigenface is applied in the process of face recognition in every class. Finally, fusion decision is adopted to get the final result of face recognition based on the intermediate result of every class. It also, can get the result of its pose recognition and its computation time is far less than that of the traditional methods. The algorithm discussed above has been tested in the experiments of a multi-pose face database with total 100 subjects and 1200 face images, which consists of ORL, UMIST, Stirling and face database of authors lab. The results are encouraging. -
R. Chellappa, C. L. Wilson, S. Sirohey, Human and machine recognition of faces: A survey, Proc.IEEE., 1995, 83(5), 705-740.[2]A. Samal, P. A. Iyengar, Automatic recognition and analysis of human faces and facial expressions:A survey, Pattern Recognition, 1992, 25(1), 65-77.[3]D. Valentin, H. Abdi, A. J. OToole, G. W. Cottrel, Connectionist model of face processing: A survey, Pattern Recognition, 1994, 27(9), 1209-1230.[4]R. Brunelli, T. Poggio, Face recognition: Feature versus templates, IEEE Trans. on PAMI., 1993,PAMI-15(10), 1042-1052.[5]S. ranganath, Face recognition using transform features and neural networks, Pattern Recognition,1997, 30(10), 1615-1622.[6]Lam Kin-Man, Yan Hong, An analytical-to-holistic approach for face recognition based on a single frontal views, IEEE Trans. on PAMI., 1998, PAMI-20(7), 673-686.[7]M. Turk, A. Pentland, Eigenfaces for recognition, Journal of Cognitive Neuroscience, 1991, 3(1),71-86.[8]A. Pentland.[J].B. Moghaddam, T. Starner, View-based and modular eigenspaces for face recognition, IEEE Conf. CVPR94, Seattle.1994,:-[9]K.M. Lam.[J].H. Yan, An improved method for locating and extracting the eye in human face images, ICPR96, Vienna, Austria.1996,:-[10]X. Xie, R. Sudhaker, H. Zhuang, On improving eye feature extraction using deformable templates,Pattern Recognition, 1994, 27(6), 791-799. -
計(jì)量
- 文章訪問(wèn)數(shù): 2221
- HTML全文瀏覽量: 151
- PDF下載量: 629
- 被引次數(shù): 0
下載:
