Research on a novel non-rigid registration for medical image based on SURF and APSO

• Published in: 2010 3rd International Congress on Image and Signal Processing Vol. 6; pp. 2628 - 2633
• Main Authors: Wang, Anna, Zhe Wang, Dan Lv, Zhizhen Fang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2010
• Subjects: affine transformation; APSO; Biomedical imaging; Convergence; Feature extraction; Image registration; non-rigid image registration; Optimization; Positron emission tomography; Robustness; SURF
• ISBN: 1424465133; 9781424465132
• DOI: 10.1109/CISP.2010.5648148

Registration of medical image has become an important research subject. For the large image distortion of some special places, the non-rigid registration is more accurate and necessary to afford important reference value for clinical care compared with the rigid registration. In order to realize robust and efficient registration, a novel non-rigid registration method for medical image based on Speeded Up Robust Features (SURF) and Adapted Particle Swarm Optimization (APSO) algorithm was proposed in the paper. In the process of registration, SURF descriptor is used to extract scale and rotation invariant features with little calculating time, furthermore APSO algorithm can avoid falling into local optimum with fast convergence speed during searching for optimal parameters. The experimental results show that the method with SURF and APSO is faster and more robust than traditional registration algorithm.