Implementation of Backpropagation Neural Network and Blood Cells Imagery Extraction for Acute Leukemia Classification
Faisal Asadi, Fiqhri Mulianda Putra, Mutik Indah Sakinatunnisa, Fadhilah Syafria, Okfalisa, Ismail Marzuki · 2017
This paper proposes an implementation of classification of Acute Leukemia using backpropagation neural network algorithm and blood cells imagery extraction. Leukemia is a cancer of blood cells, known as an abnormal white blood cells growth produced by the bone marrow. The exact cause of leukemia is still unknown. However, leukemias have been acknowledged to be grouped by how quickly the disease develops (acute leukemia and chronic leukemia) as well as by the type of blood cell that is affected (lymphocytes or myelocytes). This paper focuses on the acute leukemia, which can be categorized into Acute Lymphoblastic Leukemia (ALL) and Acute Myelogenous Leukemia (AML). These types of leukemia are possible to be diagnosed by counting the number of blood cells growth in the bone marrow through the microscopic analysis of blood cell imagery. However, it cost overpriced in terms of time, energy, and amount. In addition, manual counting may lead potential false of diagnoses. In this paper, backpropagation neural network algorithm is used to extract the characteristics of ALL and AML blood cells. Digital image processing is employed for identification type of leukemias. The experimental results argue that the proposed work achieves about 86.66% accuracy on average in classifying the leukemia acute types.