Efficient Image Matching Algorithm Using Distance Transform and Particle Swarm Optimization
Correspondence estimation is one of the fundamental challenges in computer vision lying in the core of many problems, from stereo and motion analysis to object recognition. The direct matching method would computer the fitness for every pixel in the resolution space and it would sensitive to illumination change and other change. For the image attained by distance transformation can weaken the bad influence under geometric distortion and edge change, and have the ability to resistant inversion. The PSO algorithm supports parallel search of multiple points that are changed along a smooth trajectory within the search space. The paper applied distance transformation and PSO to image matching and the experiment results showed that comparing with the traditional direct matching, the method of the paper is not sensitive to illumination change and drastically reduced the required computation time.
Y. F. Guo and H. Kai, "Efficient Image Matching Algorithm Using Distance Transform and Particle Swarm Optimization", Advanced Materials Research, Vols. 341-342, pp. 753-757, 2012