Three-dimensional (3D) acquisition technologies (for example, 3D scanners), medical visualization, and computational simulations in science and engineering can easily produce massive surface meshes with billions of triangles. The problem is that these huge meshes may not always fit into the computer’s main memory.
Well-known out-of-core mesh processing algorithms have been developed in the past decade to fix this problem. For example, to simplify an out-of-core mesh, one often segments it into multiple pieces, each of which is simplified individually. In a way, this is similar to external sorting algorithms for databases that use buffering techniques to sort large datasets on disk. In this sense, “out-of-core” means “external.”
Like most mesh simplification and refinement algorithms, a remeshing algorithm is also a multiresolution processing algorithm. Obviously, the idea of remeshing is not new in geometry processing, even for large meshes, but the authors were able to step forward in designing and implementing a remeshing algorithm for huge meshes.
Nevertheless, the main contribution of this paper is that the remeshing algorithm implemented in a streaming framework runs by synchronously streaming the input and the simplified mesh of the same surface. This synchronization strategy allows us to establish a correspondence between a simplified mesh and a huge input mesh of the same surface, which is needed for many applications and editing operations.
To conclude, this paper is particularly useful for those who are faced with the increasing size of geometry datasets, in the quest for realistic scenarios in many knowledge areas of science, engineering, arts, and entertainment computing.