Layered Recognition Scheme for Robust Human Facial Expression Recognition using modified Hidden Markov Model

Facial Expression Recognition is an application used for biometric software that can be used to recognize special expressions in a digital image by analysing and comparing different patterns. These software are popularly used for the purpose of security and are commonly used in other applications such as credit card verification, surveillance systems, medicines, home security, human-computer interface, etc.. Recognizing faces becomes very difficult when there is a frequent change occurs in facial expressions. In this paper two layer extension of HMM is used to identify continuous effective facial expressions. Partition technique is used for feature extraction. Two layered extension of HMM consists of bottom layer which represents the atomic expression made by eyes, nose and lips. Further upper layer represents the combination of these atomic expressions such as smile, fear etc. In HMM, Baum-Welch, Viterbi and Forward methods are used for parameter estimation for calculating the optimal state sequence and probability of the observed sequence respectively. This proposed system consists of three level of classification. Output of the first level is used for the training purposes for the second level and further this level is used for the third level for testing. Six basic facial expressions are recognised i.e. anger, disgust, fear, joy, sadness and surprise. Experimental result shows that Proposed System performs better than normal HMM and has the overall accuracy of 85% using JAFFE database.