Title :
Algorithm of thermal image segmentation based on Gaussian mixture model and artificial fish swarm algorithm
Author_Institution :
Faculty of Automation, Guangdong University of Technology, Guangzhou, China
Abstract :
The accurate thermal image segmentation becomes extremely challenging in the complex background conditions, especially in the circumstances that the foreground target and the background are very similar. The paper uses the Gaussian mixture model (GMM) for the rough segmentation in the temporal domain to solve the problem of complex background, then for the fine segmentation in the spatial domain by using 2d fuzzy partition maximum entropy method in order to overcome the difficulties of accurate segmentation in the case of the foreground target and background are very similar, and introduces the artificial fish swarm algorithm (AFSA) to search the combined parameters of segmentation quickly. Simulation results show that this algorithm is robust, good real-time, and achieved good segmentation effect.
Keywords :
Biomedical imaging; Entropy; Image segmentation; Integrated circuits; RNA; Robustness; Artificial Fish Swarm Algorithm(AFSA); Gaussian Mixture Model (GMM); Thermal image segmentation;
Conference_Titel :
Conference Anthology, IEEE
Conference_Location :
China
DOI :
10.1109/ANTHOLOGY.2013.6784906