Evaluating the effects of K-means clustering approach on medical images

Image segmentation is an essential process for most analysis tasks of medical images. That's because having good segmentation results is useful for both physicians and patients via providing important information for surgical planning and early disease detection. This paper aims at evaluating the performance of the K-means clustering algorithm. To achieve this, we applied the K-means approach on different medical images including liver CT and breast MRI images. Experimental results obtained show that the overall segmentation accuracy offered by the K-means approach is high compared to segmentation accuracy by the well-known normalized cuts segmentation approach.