Hybrid Multi-Objective Active Appearance Model for Gamer’s Facial Features Detection

In this article we propose a robust facial analysis of multiple camera system for cyber games which is capable of extracting the facial features of a face making large lateral movements. Tactical maneuvers of the gamer make single camera acquisition system unsuitable to analyse and track the face due to his large lateral movements. Although our proposition of double camera acquisition system resolved this problem, but the facial data obtained from both cameras produces optimization problems for face search algorithms. For an improved facial analysis system, we propose to acquire the facial images from two cameras and analyse them by Pareto based hybrid multi-objective face search optimization for 2.5D Active Appearance Model (HMOAAM). Proposed algorithm is applied on number of multi-view real and synthetic facial images and its results are compared with a non hybrid system. Results obtained validate our proposition.