Human Spermatozoa Micrographs Image Segmentation and Classification to Assess Dysspermia

Abstract

The medical imaging field has advanced considerably over the last few decades and new techniques such as Photomicrograph, Electron micrograph and Digital micrograph are able to obtain the spermatozoa images in noninvasive ways. These new technologies have opened up new opportunities for us to explore novel applications to determine the sperm count, motility and morphology. This thesis presents research and development of methods for three major computational issues in sperm imaging research: detection, segmentation, tracking and classification. Three categories of features are extracted from spermatozoa micrograph images: First Order Statistics (FOS), Gray Level Co-occurrence Matrices (GLCM) and Morphological features. Principal Component Analysis (PCA), Singular Value Decomposition (SVD) and Factor Analysis (FA) are employed to reduce the number the extracted features in order to improve the accuracy of classifiers. The performance of PCA SVD and FA in feature dimensionality reduction is investigated. newline