Simplex particle swarm optimization for block matching algorithm

In this paper, a novel block matching algorithm SPSO based on hybrid simplex method (SM) and particle swarm optimization (PSO) is proposed, which combines the high accurate local search ability of SM and the powerful global search ability of PSO. To reduce computational complexity and improve the performance of block matching algorithm, the selection of initial habitats based on fixed points and random points, the early termination and the SM searching after finishing each iterating of PSO are used. Experimental results verify that the SPSO algorithm can achieve better PSNR value than conventional block matching algorithms TSS and DS with little sacrifice of time consumption, especially for violent motion.