A method used in myeloma cell cloning spot image segmentation.
- Author:
Yu-ya YAO
1
;
Zhuang-zhi YAN
;
Yu CHEN
;
Shu-peng LIU
Author Information
1. School of Communication and Information Engineering, Shanghai University Institute of Biomedical Engineering.
- Publication Type:Journal Article
- MeSH:
Humans;
Image Enhancement;
methods;
Image Interpretation, Computer-Assisted;
methods;
Multiple Myeloma;
classification;
Pattern Recognition, Automated;
methods;
Tumor Stem Cell Assay;
instrumentation;
methods
- From:
Chinese Journal of Medical Instrumentation
2008;32(5):319-322
- CountryChina
- Language:Chinese
-
Abstract:
A new interactive image segmentation method used in the multiple myeloma cloning spots image segmentation is presented in the paper. Based on the theory of graph cuts, some pixels are selected as the front object and the background seeds, and the other parts are treated as the unknown region. Then, an energy function is constructed and initialized through K-means, and the minimum cut method is used in the segmentation by energy minimization. Last, the image is eroded and dilated, and the cloning separate parts could be got effectively. For the pixels which may be partitioned wrongly, we use a tool similar to a brush to re-mark the front object or the background, and divide once again. Both subjective the evaluation criteria and the RUMA, evaluation criteria are used to evaluate the method, and the experiment results are satisfactory.