Interval graph of facial regions with common intersection salient points for identifying and classifying facial expression

Measuring the facial expression is an important research and used in many real-time applications. Various methods are proposed in the academia and industry for a decade and still continue to have research potential. This paper proposes a novel scheme by using Interval graph of facial regions. It is assumed that common intersecting salient points of facial regions can be used for estimating the emotions. The facial region is decomposed in four sub regions and the Interval graph is extracted for each region. The common salient points and degree of deformation and direction of deformation are measured for vertical, horizontal and diagonal directions. These values are considered as feature vectors. The well-known datasets such as JAFEE and CK++ are used for evaluating the performance of various classification algorithms and estimating their average classification accuracy. The average classification of the proposed approach is 95.9% and 94.7% for CK++ dataset and JAFEE dataset respectively. The performance of the proposed approach is better when compared to other state of art approaches.