Classification based on Neural Networks to Detect Acute Lymphoblastic Leukemia

Acute Lymphoblastic Leukemia (ALL) can be detected using Artificial Intelligence (AI) techniques. For this purpose, the images captured by a microscope of human peripheral blood smear samples are analyzed and recognized. This article presents an automated method for identifying and classifying white blood cells using special neural networks. The image database of ALL has been used. We focused on extracting texture characteristics (histogram of brightness, contrast, and orientation of contours). These characteristics are presented to the input of the Random Threshold Classifier (RTC). The input encoder generates the output binary vector, which is then presented to the RTC neural classifier, and, finally, the classifier's output provides the recognized class. In this case, we have three classes, healthy cells, affected cells, and background. In this work, we compare the performance of two classifiers, the RTC and the Limited Receptive Area Grayscale classifier (LIRA Grayscale). As shown below, the RTC neural classifier achieved a recognition rate of 98.3%, while the LIRA classifier achieved a rate of 96.56%. Our system is evaluated using the public dataset of peripheral blood samples from ALL-IDB.