The Community for Technology Leaders
Green Image
This paper presents a hybrid framework of feature extraction and hidden Markov modeling(HMM) for two-dimensional pattern recognition. Importantly, we explore a new discriminative trainingcriterion to assure model compactness and discriminability. This criterion is derived from hypothesis testtheory via maximizing the confidence of accepting the hypothesis that observations are from target HMMstates rather than competing HMM states. Accordingly, we develop the maximum confidence hiddenMarkov modeling (MC-HMM) for face recognition. Under this framework, we merge a transformationmatrix to extract discriminative facial features. The closed-form solutions to continuous-density HMMparameters are formulated. Attractively, the hybrid MC-HMM parameters are estimated under the samecriterion and converged through the expectation-maximization procedure. From the experiments onFERET and GTFD facial databases, we find that the proposed method obtains robust segmentation inpresence of different facial expressions, orientations, etc. In comparison with maximum likelihood andminimum classification error HMMs, the proposed MC-HMM achieves higher recognition accuracieswith lower feature dimensions.
Parameter learning, Statistical, Classifier design and evaluation, Face and gesture recognition
Jen-Tzung Chien, Chih-Pin Liao, "Maximum Confidence Hidden Markov Modeling for Face Recognition", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 30, no. , pp. 606-616, April 2008, doi:10.1109/TPAMI.2007.70715
47 ms
(Ver 3.3 (11022016))