Circular Hough Transform and Region Based Methods for Counting of Red Blood and White Blood Cells.

Abstract

Cells in the blood are major indicators of health. The number of cells in blood indicates the underlying illness and it will help to identify the source of illness. Blood cells are major biomarkers present in human body. Haemocytometer is the common and basic device which is used to count the blood cells. By appropriate staining procedures, we shall probably count different cells in the blood. RBC is stained using Haem’s solution and WBC is stained using Turk’s solution. This works involves the automatic counting of blood cells by merging the staining results with computer. The stained cells are captured with microscope and the same is interfaced with computer using ultra scope. The ultra scope output is called in MATLAB and major preprocessing steps like re sizing, RGB to Gray, histogram equalization is done. RBC image is then segmented based on shape using region based segmentation method and overlapped RBC’s are identified by circular Hough transform. The WBC count is identified by subjecting WBC image to edge and region based segmentation methods. The segmentation algorithm will provide the count within given mm of blood. This method helps in avoiding man-made error that occurs during manual counting of cells using haemocytometer. Thus, the proposed method can be implemented to construct home based identification equipments.