Development of Anemia Cells Recognition System Using Raspberry Pi

Abstract

Anemia is the most predominant blood disease globally and is caused by iron deficiency resulting to fatigue. Thalassemia is the shortage of production of essential protein, the hemoglobin, in the blood that distribute oxygen throughout our body. Codocytes or Target cells and Elliptocytes are types abnormal red blood cells that are most commonly associated with anemia and Thalassemia. Traditional method of manually determining these abnormal RBCs from a blood smear is labor intensive and can be subjective. This paper automates the recognition of codocytes and elliptocytes from blood smear images. The recognition system uses image processing and support vector machine to be able to classify the Codocytes and Elliptocytes in the PBS. The average accuracy for the classification of PBS images that contain codocytes and elliptocytes is 94.31%. This will help advance further researches on abnormal red blood cell detections and aid in identifying early pathognomonic determinants of anemia and Thalassemia.