Fast Covariance Matching Based on Genetic Algorithm

This paper proposes an effective framework to boost the efficiency of covariance matching. In this framework, covariance matrices are used to match object in complex environment by fusing multiple features. Then, Genetic Algorithm (GA) is employed to improve the processing speed of covariance matching. To take advantage of the property of GA for the optimization in large search spaces to covariance matching, a fitness function is designed using the distances between the covariance matrices of model and candidate regions. Experimental results show that the proposed approach can improve the processing speed of covariance matching observably. The computing speed of the proposed method is at least 7 times than that of exhaustive searching.