Segmentation of Human Spermatozoa using Threshold-Based Image Segmentation

Abstract

The role of image processing in processing and analyzing the microscopic medical images is the most challenging and required task in the assisted method of fertilization for human society. The Human eye evaluation for the process of detecting the defective spermatozoa from the sample semen smear using the microscope yields subjective results, which may vary from person to person. The objective evaluation is based on an automated computer program segments the portion of interest from the image based on segmentation techniques. The effective segmentation in the medical image is to highlight the expected portions such as head, tail, and mid-piece for the further process of analyzing the defects in the sperm cell. Cluster-based image segmentation is one of the effective methods to segment the object from the background in the microscopic medical images [1]. Entropic thresholding techniques also had an impact on the segmentation of medical images [2]. We have implemented the various threshold based image segmentation and compared their results with the help of segmentation metrics and showed the effectiveness of thresholding techniques for microscopic medical images.