Application of digital image processing is increasingly popular in medical field to process medical image towards increasing the accuracy and efficiency in medical image diagnosis. In this project, a study on the suitability of mean shift to perform clustering and segmentation on medical images is conducted. Mean shift is an iterative procedure that shifts each data point to the average of data points in its neighborhood. Experimental data in this study are pap smear images, thin blood smear images, mammograms and ultrasound images. The proposed segmentation algorithm commence with a filtering step as a pre-processing method, followed by clustering step and post-processing step. The main aim of the project is to study the performance of mean shift clustering as compared to k-means, moving k-means and fuzzy c-means clustering. For filtering, mean shift filtering is compared with conventional median filtering in image noise filtering. Computer logic is used as a feature counting technique to count the region of interest (ROI) in the clustered image. Experimental result shows mean shift filter performs better than median filter in noise reduction. For clustering and segmentation, result shows mean shift clustering has higher capability to cluster and segment the ROI accurately as compared to k-means, moving k-means and fuzzy c-means clustering due to the adaptive clustering nature of mean shift. As a conclusion, mean shift clustering algorithm is suitable and capable to cluster and segment medical images.