A comparative study between methodologies based on the hough transform and watershed transform on the blood cell count

Abstract

It is increasingly common to use engineering techniques in the areas of health, in order to solve simple problems or even create new diagnostic methods. In the last decade, the Hough Transform has been widely used as a tool for segmentation of blood smear images for the purpose of counting blood cells. However, it is noted that the Watershed transform has been applied to perform the same function. Based on this, a methodology based on the Hough Transform was created, aiming to perform the detection and counting of erythrocytes and leukocytes and verify the applicability of the methodology when compared to others. The study was conducted based on the determination of accuracy and simulations performed on different hardware platforms and subsequent comparison with the WT-MO methodology. The results demonstrated that both methodologies are able to perform the task of detection and counting of blood cells in digital images of blood smear. However, the methodology based on the Watershed Transform best meets the criteria of speed and reliability (counts), which are indispensable to medical laboratory routine.