Recognition of Facial Expression Used Spatial Frequency and Self-Organizing Map for Human Interaction

Abstract

In this research, we aim to build the system which is able to recognize the facial expression that expresses emotions most directly, by using spatial frequency and self-organizing maps (SOM). In this method, we use the spatial frequency which is able to recognize the shape change of face organ like eyes and mouth as an input data to SOM which is able to visualize high dimension data. Spatial Frequency is able to obtain the subtle shape changes of face organ as the frequency component and compress the information of image. We propose a method which is able to recognize the expression with higher precision by extracting the necessary shape information from the face image by using the spatial frequency.