Utilizing Structure-rich Features to improve Clustering