Automated RBC Morphology Counting and Grading Using Image Processing and Support Vector Machine
Rosemarie V. Pellegrino, Aubrey C. Tarrobago, Dave Lester B. Zulueta
Abstract
Red Blood Cell (RBC) morphology such as Target Cells and Elliptocytes characterize early pathognomonic determinants in certain diseases like Iron Deficiency Anemia and Thalassemia. Significant amounts of target cells or elliptocytes in a blood sample can be used to grade the existence of Blood Related Disease. In the Philippines, 37.6% of Filipinos have Iron Deficiency Anemia (IDA) and 27.8% suffer from Thalassemia. This study automates the classification, counting, and grading of RBC morphology using image processing techniques and SVM classification. The researchers acquired PBS samples and designed a prototype capable of analyzing these with a Raspberry Pi computer. The device classified, counted, graded and provided associated disease considerations of the sample PBS test. Comparison of the machine and hematologist’s reading of the normal red blood cells, target cells and elliptocytes samples gave an average accuracy of 95.77%.