Analysis of Image Segmentation Algorithms for the Effective Detection of Leukemic Cells

Abstract

Image segmentation plays a vital role in medical image processing. Different pre-processing methods yield different results. The pre-processing methods such as histogram stretching with erosion and dilation, average filter and median filter along with histogram stretching is applied to the four different segmentation algorithms which are Otsu's thresholding, Watershed based segmentation, Canny edge detection and K-mean clustering. These algorithms are used to segment Acute Lymphoblastic Leukemia datasets and the parameters such as precision, accuracy and sensitivity of the results are calculated so as to find a better algorithm which is suitable for segmentation of the leukemic cells.