Medical Image Registration Based on an Improved Ant Colony Optimization Algorithm
Ting Xun Lin and Herng Hua Chang
Computational Biomedical Engineering Laboratory (CBEL), Department of Engineering Science and Ocean Engineering, National Taiwan University, Daan 10617 Taipei, Taiwan
Abstract—Image registration is one of the fundamental and essential tasks within image processing. It is the process of determining the correspondence between structures in two images, which are called the template image and the reference image, respectively. The challenge of registration is to find an optimal geometric transformation between corresponding image data. This paper develops a new image registration algorithm that is based on an improved ant colony optimization algorithm. In our approach, the image pixels are treated as the nest of a swarm of ants. The ants are designed to have the ability to forage for the “food” in their memory. Subsequently, the ants deposit pheromone on the pixels, which affect the motion of the ants. The registration process of updating the pheromone, the direction and distance of advancement is repeated until the correlation coefficient between the registered and reference images reaches a maximum. Experimental results indicate that our method accurately transformed the template images into reference images in various scenarios. It is indicated that the proposed method is of potential in a wide variety of image registration applications.
Index Terms—ant colony algorithm, image registration, transition probability
Cite: Ting Xun Lin and Herng Hua Chang, "Medical Image Registration Based on an Improved Ant Colony Optimization Algorithm," International Journal of Pharma Medicine and Biological Sciences, Vol. 5, No. 1, pp. 17-22, January 2016. doi: 10.18178/ijpmbs.5.1.17-22