EPSMS for acute leukemia classification
This section describes our proposed approach for acute leukemia classification using EPSMS. We first describe the scenario. Next, we describe the methods used for cell segmentation and feature extraction. Finally, we describe how EPSMS was used to select competitive classification models.
Morphological identification of leukemia Our problem was the morphological classification of acute leukemia from digitized bone marrow images. The scenario can be summarized as follows. First, digitized images were obtained. After preprocessing the images (to adjust for the contrast and filtering noise), the first step of the methodology was image segmentation, which involved the identification of regions of interest in images. Features were then extracted from the regions identified, before
the classification model was built. Figure 2.2 shows the scenario we considered. In this study,weused a database of cell images from real patients [10]. In each record in