Facial Expression Generated Automatically Using Triangular Segmentation

Abstract

The advancement in computer technology provided instant messaging software that makes human interactions possible and dynamic. However, such software cannot convey actual emotions and lack a realistic depiction of feelings. Instant messaging will be more interesting if users’ facial images are integrated into a virtual portrait that can automatically create images with different expressions. This study uses triangular segmentation to generate facial expressions. The application of an image editing technique is introduced to automatically create images with expressions from an expressionless facial image. The probable facial regions are separated from the background of the facial image through skin segmentation and noise filtration morphology. The control points of feature shapes are marked on the image to create facial expressions. Triangular segmentation, image correction, and image interpolation technique are applied. Image processing technology is also used to transform the space of features, thus generating a new expression.