Segmentation of TB Bacilli in Ziehl-Neelsen Sputum Slide Images using k-means Clustering Technique

Raof, Rafikha (Universiti Malaysia Perlis, Pauh Putra, Perlis, Malaysia)
Mashor, M. Y. (Universiti Malaysia Perlis, Pauh Putra, Perlis)
Noor, S. S. M. (Universiti Sains Malaysia, Kubang Kerian, Kelantan, Malaysia)

CSRID (Computer Science Research and Its Development Journal)
Vol 9, No 2 (2017): CSRID Juni 2017

Image segmentation is the most crucial steps in determining the accuracy of a medical diagnosis system that is based on image processing procedures. Therefore, it is important to select a suitable image segmentation technique to obtain good results and hence providing optimum accuracy for the developed diagnostic system. In this research, image segmentation procedure using k-means clustering approach has been considered for differentiating between pixels that represent TB bacilli and pixels that represents sputum or background. This paper presents the technique used to separate the TB bacilli and its background from the Ziehl-Neelsen sputum slide images. The k-means clustering has been applied to those images followed by several extra rules. The resulted images show encouraging results, which indicate that the proposed segmentation method is able to filter out the TB bacilli pixels from the background pixels.