Interest Points Detection of 3D Mesh Model Using K Means and Shape Curvature

Due to the great improvement of technology, the representation of the data in three dimensions widely used in many applications like scientific visualization, manufacturing, computer vision, engineering design, virtual reality, architectural walk through, and video gaming. The 3D objects consist of a huge number of components like thousands of vertices and faces, and that make researchers to detect the more interest components to deal with this objects instead of to deal with the whole object that for many applications like a watermark, and mesh simplification. This paper presents a model to detect the interest points of the object using k mean clustering algorithm depend on the curvature areas. The proposed model is applied to a set of benchmark data of 3D mesh models. The evaluation of the proposed model is verified using three evaluation measures, namely False Positive, False Negative Errors, and Weighted Miss Error. The evaluation is also verified by comparing the proposed model with the most popular interest point detecting techniques.