A new approach for overlapping cell separation in pre-processed sickle images using SCP model.

Identification of sickle cell disease (SCD) plays a crucial role in healthcare image analysis. By categorizing the red blood cell abnormalities and cell counts, it comprehensively analyzes the illness and provides an accurate identification. The diagnosis of SCD severity for the diagnostic procedure is dependent significantly on the detection process. Cell morphology performs an important role in distinguishing between normal and sickle cells. The analysis of the cells is complicated by the existence of overlapping cells in the images. The proper segmentation of overlapping cells is essential for SCD identification. This study examines how traditional image processing and a deep learning model separate overlapping cell. Before segmentation, images have been enhanced and pre-processed to ensure that the overlapped cells are accurately separated. We evaluate the deep learning model of the Star Convex Polygon method (SCP) with pre-processed images and raw images to investigate how the image pre-processing technique affects overlap cell segmentation. We analyzed 80 images to test and achieved 97.96% accuracy for overlap cell segmentation using the SCP model and pre-processed images. [ABSTRACT FROM AUTHOR]