Sun Nongliang, Xu Saicong, Cao Maoyong, Li Jing
Member, IEEE, College of Information and Electrical Engineering, Shandong University of Science and Technology, Qingdao, China. (Tel:+86-532-86057928; fax: +86-532-86057153; e-mail:
Conf Proc IEEE Eng Med Biol Soc. 2005;2005:3324-7. doi: 10.1109/IEMBS.2005.1617188.
Correctly counting the live cells plays a great role in the ectogenetic anti-virus experiment. According to the irregular shape and arbitrary size of the wall pasted Hela cells overlapping each other, we propose a scheme to segment and count the cells using Gabor filter with different parameters and Morphological operation. Experiments reveal that filters with different parameters will lead to different results and a better segmentation will be achieved based on the characteristics of cells and optimal parameters. Large amount of experiment results show that this algorithm can successfully segment the cells and the accuracy arrives at 99.3%. This scheme based on image analysis and pattern recognition can overcome some disadvantages of traditional approaches, shortening anti-virus experimental period and reducing experimental cost.
