I know when you are happy - emotion detection

Abstract

Affective computing is an interdisciplinary research field. Various approaches have been proposed for the detection of emotions. One of the approaches is emotion detection from facial expressions. People react to situations in real life and they have little or limited control over their facial expressions. These facial expressions can be examined to know about their emotional state. In this paper we propose an approach based on Laplacian of Gaussian filters for frontal facial images. This technique removes the disadvantages of Laplcian filters susceptibility to noise. Also the simplicity of these filters makes them eligible for timely computations. For emotion recognition all possible two class emotion combinations are used. Experiments are conducted on expression dataset available for research purpose.