A Study on Segmentation and Evaluation of Leukocyte Section

Abstract

Clinical image assessment plays an imperative part in the assessment of the illness in humans. Hence, in most of the disease cases, a suitable procedure is followed to record the bio-medical image in a controlled environment, which further can be evaluated by an experienced doctor or a suitable Computed-Assisted-Diagnosis (CAD) system to confirm the severity of disease; for planning suitable treatment process. In this work, a morphological operation based CAD system is developed to examine the leukocyte segment from the RGB-scale blood smear image. The presentation of developed CAD is evaluated with the other leukocyte examination procedures existing in the literature. The outcome of this work substantiate that, proposed approach is fast and efficient in extracting the leukocyte section from the blood-smear picture with improved accuracy. In this work, the validation of the proposed technique is experimentally authorized with the benchmark LISC database and the results obtained with the proposed study offers an overall accuracy of > 97.96% compared to the related existing techniques.