Abstract: We present an algorithm for generating point distributions with high-quality blue noise characteristics based on centroidal Delaunay triangulation. The method combines Delaunay triangulation with centroidal patch triangulation, and iteratively moves each vertex to the centroid of its 1-ring neighborhood and updates the topological connectivity between the sampling points. The centroid of a patch is calculated by using a given density function. The experimental results demonstrate the effectiveness and robustness of the proposed algorithm.