Robust and efficient approach to diagnose sickle cell anemia in blood

Abstract

Red Blood Corpuscles (RBCs) form the main cellular component of human blood. RBCs in common physiological status has a circular form in the front view and bi-concave form in side view. In case of a person is infected with anemia RBCs form as sickle-shaped cells which drive to blood vessel obstruction joined by painful episodes and even death. Precise and robust cells classification and counting are essential in evaluating the level of anemia disease danger. The classification and counting of red blood cells (normal and abnormal RBC cells) are challenging due to the complex and heterogeneous shapes and overlapped cells. In this paper, we propose a new robust approach to classify red blood cells to two groups: Normal and Abnormal RBC cells based on area and Eccentricity of each cell, then count the total number of normal and abnormal RBC cells individually. For the sake of comparison, we also implement the latest Sickle cell research, which uses circular Hough transform. We compare our approach to circular Hough transform in the same execution environment. Our new approach sets the state-of-the-art performance in term of effectiveness (cell counting) and efficiency (execution time).