An effectual facial expression recognition using HMM

In general human face has similar and different characteristics which play a very important role in recognizing facial expression. In this work, a new method presents to recognize different facial expressions from time sequential facial expression images. The performance of an automatic facial expression recognition system can be significantly improved by modeling the accuracy of various streams of facial expression information utilizing multi stream hidden Markov models (HMMs). Active Appearance Model (AAM) landmarks are measured for each frame of the videos. The AAMs were used to identify the face and extract its graphic features. HMM is act as a classifier for the expression prediction. The k-NN used for classification or regression because of its lazy learning also able to identify the person in the image. The experiment explored that the proposed approach has show encouraging accuracy in realizing all expression.