Proficient facial expression recognition using HLACLF extraction with a bank of Bayesian classifiers for recognition

This paper presents a proficient method for extracting the HLAC-like features. Different masks from 2D-monochrome image sequences will be used for extracting the features. The Mutual Information Quotient (MIQ) is the basis for selecting the most relevant features. In this study, a bank of Bayesian classifier is adopted for recognition. The distinguished features are given to a bank of seven parallel Bayesian classifiers. For recognizing a particular facial expression, each Bayesian classifier is trained. The outputs of all classifiers are then combined using a maximum function. The test will be performed on images from the JAFFE database.