Mesh Simplification Algorithm Based On Kmeans

Grid clustering model in computer graphics applications field such as computer animation, interactive visualization and virtual reality has broad prospect. The existing grid clustering algorithm may be spend more time in complex, but the result is not the perfect. Especially the data point clustering in accurate phenomenon at the edge of the clusters. This paper proposes a new clustering algorithm based on face clustering, In order to getting the sampling points, the algorithm utilizes the most far geodesic distance adaptive divided. Based on the sampling points, we utilizes the function of measurement to clustering triangle mesh, and then take advantage of PCA algorithm and KMeans algorithm make the cluster more compact and regular. Without changing the original model data, the complex computing time greatly reduced. The algorithm is simple, fast and good to keep the boundary feature. A set of illustrations are given to show the effectiveness of the algorithm.