Designing A Two-Stage Classification Network Algorithm for Acute Lymphocytic Leukemia Diagnosis in Blood Lamella Images

Diagnosis of leukemia is very difficult, therefore, it is necessary to use image processing techniques. The main objective of this study was to provide a system based on intelligent models that could improve the accuracy of the diagnostic system for acute leukemia. The images produced in this study were extracted from the University Degli Studi Dimilan database and processed in the MATlab 2014a software. In this research, Fuzzy-Cmeans method was used in fragmentation and neural network and support vector machine in classification networks. Ethical Considerations: In this study, all principles of research ethics were considered. Feature data were extracted using the original image transfer to RGB, HSV, Lab and Enhanced RGB spaces. The data obtained from the previous step were entered into the SVM network, then the network separated normal data from abnormal data. The results of comparing the output of the proposed method with different educational methods showed the highest mean of accuracy equal to 95.7%. The application of the proposed network in this study was that eliminate the weak points of all the networks in addition to presenting the advantages of these network. Combining the networks improved the accuracy of output up to 98% and considerably reduced the time required for calculations.