جلال الدین نصیری - Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
هادی صدوی یزدی - Department of Electrical Engineering, Tarbiat Moallem University of Sabzevar, Sabzevar, Iran
محمود نقیب زاده - Commutation and Computer Research Center

Finding mouth in face is a bottleneck of many applications such as face detection and lip reading. In this paper explain a new approach for mouth detection using Particle Swarm Optimization (PSO). PSO is used to mining the rule of between pixels of mouth and other pixels an optimized map. The image is mapped to YCbCr color space. The main idea of the method is based on that Mouth has the high values of Cr and low values of Cb. The proposed algorithm has been examined on CVL and Iranian databases and we have reached to the 92% correction rate which comparing to the previous approach, there is 11% increase in Mouth detection. We find out that the proposed algorithm is flexible which it is independent of lightening conditions.

Mouth detection, PSO, Color images,YCbCr Color Space