Acute lymphoblastic leukemia image segmentation based on modified HSV model

Abstract Image segmentation is a critical step in computer-aided diagnosis that could speed up Leukemia detection. cancer of the blood has reputation for being particularly lethal. Based on immunohistochemical method, leukocytes can be manually counted stained peripheral smear image to detect Acute Lymphoblastic (ALL). Regrettably, manual process takes about 3 24 hours complete, which insufficient. This paper introduced new and straightforward ALL approach based color transformation. First, Leukemia, ALL-IDB1, ALL-IDB2, datasets were used this paper. The dataset includes 208 ALL-IDB1 ALL-IDB2 images, while 3256 images. Next, we use HSV model transform In addition, modified by pre-processing saturation channel better results. Then, pre-processed images segmented fixed threshold. After that, various metrics are utilized measure output proposed method. Finally, methodology compared currently benchmarks. method outperforms previous approaches regarding accuracy, specificity, sensitivity, time. results show technique improves performance measures significantly.