Firefly Algorithm-Based Kapur’s Thresholding and Hough Transform to Extract Leukocyte Section from Hematological Images

Computerized disease examination techniques are widely adopted in the literature to evaluate a considerable number of medical images ranging from the RGB scale to gray scale. This work proposes a novel image extraction method by combining Kapur’s thresholding and Hough transform (HT) to extract the leukocyte segment from the RGB-scaled blood smear image (BSI). Automated mining of the leukocyte region is always preferred in medical clinics for fast disease examination and treatment for the planning process. This study aims to implement a hybrid procedure to extort the leukocyte segment. Kapur’s is considered to enhance the RGB-scaled test image, and the HT is used to detect and extract the circle section from the image. In this work, the hematological images of leukocyte images for segmentation and classification (LISC) database are adopted for the examination. The extracted leukocyte picture is then evaluated with ground truth, and the essential image performance parameters (IPP) are then computed. This work is then validated against the semiautomated approaches, such as Shannon’s entropy-based Chan-Vese and level-set segmentation techniques existing in the literature. The outcome of the proposed techniques confirms that proposed procedure gives better IPP values compared to the existing semiautomated techniques.