Fusion 7×7 median filter and seeded region growing area extraction algorithms for effective detection of acute leukemia based on blood images

Abstract

In the most medical image analysis tasks, image segmentation is a crucial step where one of the usual segmentation methods is based on clustering algorithms. Clustering algorithms have been employed as a digital image segmentation method in various fields, including engineering, computer and mathematics. This study aims to detect acute leukemia cells based on blood images with effective methods. Normally, although the segmentation process is completed, there is still the existence of unwanted noise and background in acute leukemia blood images. Thus, this study proposes a fusion of median filter and seeded region growing area extraction algorithm to overcome this problem. Finally, the results demonstrate that the proposed method has successfully been distinguished and segmented between acute leukemia cells from the unwanted noise or complicated background in blood images. The proposed method is able to yield an average of 98.26% based on the percentage of accuracy.

Keywords: Acute leukemia, median filter, Seeded Region Growing Area Extraction, k-Means