Comparative Analysis of K-Means and K-Nearest Neighbor Image Segmentation Techniques
Prachi Surlakar, Sufola Das Chagas Silva Araujo, K. Meenakshi Sundaram · 2016
Image segmentation technique assigns a label to every pixel in an image, such that certain similar characteristics are shared by pixels with same label. Segmentation simplifies the representation of an image into something that is more meaningful and easier to analyze. It plays a crucial role in medical diagnosis and treatment of diseases by cutting out Region Of Interest (ROI) from an image. The ROI differs based on the application and thus image segmentation still remains a challenging area of research. This paper presents the comparison of K-Means and K-Nearest Neighbor image segmentation techniques for segmenting the slide of Syringocystadenoma papilliferum which is a sweat gland tumor appearing at birth or puberty. Segmentation technique is used by pathologists to distinguish different types of tissues and focus on the region of interest. The evaluated results with different algorithms showed that K-NN segmentation technique revealed higher mutual information, hence proving it to be comparatively a better algorithm.