Features Classification for the Extracted ROIs of Microscopic Pap Smears Images
E. Olusegun George, Yusra A. Sroor, Haider G. Hussein · 2013
Pap smears is a very effective screening test for cervical precancerous. However, hundreds of small windows have to be looked under microscope by a trained cytologist for a single slide from each patient. It makes this process very tedious and erroneous. The proposed system is designed to give a discriminati on ability of the adopted features to do suc- cessful classification for the extracted ROI (i.e., nucleus of the cells). Two kinds of features have been used for discriminati on purpose; the first kind is a set of geometrical features because many of the extracted Pap smear ROI are categorized ac- cording to their shapes (i.e., circular, oval, irregular, small, big). While the second kind of features is a set of texture fea- tures (local and global color features) because one of the main differences between cancerous cells and other types is the local changes in the brightness.