AUTOMATED SEGMENTATION METHODS FOR MICROSCOPIC BLOOD CELLS

摘要

In haematology, developing an automated blood cells types identifications based on microscopic examination could serve as a software tool for archiving haematology images, for automatic classification and it could help widely the clinicians in their laboratory tasks. This paper presents basic algorithms used in automated blood cell identifications based on microscopic image processing techniques. White blood cell segmentation method, red blood cell segmentation method and classification of basic blood cell types are presented. The promising results are compared with the manual investigation results obtained by the haematologists and experts of the field. A ratio of white blood cells over red blood cells is computed and gives an idea of the clinical diagnosis of the blood smear examined and all the examined samples matched the real clinical diagnostic.