A Cascaded Classification-Segmentation Reversible System for Computer-Aided Detection and Cells Counting in Microscopic Peripheral Blood Smear Basophils and Eosinophils Images

Computer-aided image analysis has a pivotal role in automated counting and classification of white blood cells (WBCs) peripheral images. Due to their different characteristics, our proposed approach is based on investigating the variations between basophils eosinophils terms color histogram, size, shape before performing segmentation process. Accordingly, we cascaded system using classification-based process, called classification-segmentation reversible (CSRS). Prior applying CSRS system, Histogram-based Object Background Disparity (HOBD) metric was deduced determine most appropriate plane for initial WBC detection (first segmentation). Investigating local histogram features both classes resulted 92.4% accuracy third-degree polynomial support vector machine (SVM) method. Subsequently, approach, transformation-based algorithms were developed fit specific requirements each two predicted classes. The used, where images from an process are fed into second class separately. results demonstrated similarity index 94.9% basophils, 94.1% eosinophils. Moreover, average 97.4% achieved. In addition, carried out after CSRS, achieving 5.2% increase compared