Article ID Journal Published Year Pages File Type
488641 Procedia Computer Science 2015 10 Pages PDF
Abstract

Overlapping cells in a medical image can affect the extraction process of cells’ characteristics as they are visually indistinguishable from each other. This paper proposes a new separation process algorithm called the radial-based cell formation (RCF) for overlapping cells. The algorithm has been tested on 160 overlapping red blood cells taken from thin blood smear images, and 26 overlapping cervical cells taken from Pap smear images. On all tested medical images, the cells of interest are successfully detected and distinguishable from those overlapping cells, with size and shape well-preserved. The separated cells have been analyzed with both qualitative and quantitative analyses, resulting in much better results compared to other conventional algorithms. The correlation test approach was then used to determine the performance of the proposed RCF algorithm compared to the manual visual interpretation by the pathologists. The results obtained prove that this study has successfully developed a new separation process algorithm (i.e. RCF) for overlapping cells in medical microscopic images, with well-preserved size and shape of the individual cell.

Related Topics
Physical Sciences and Engineering Computer Science Computer Science (General)