Deep learning and GPU-accelerated algorithms for computer-aided engineering

Deep learning frameworks for point cloud reconstruction

“…For the physical design of such models, if S represents the solid, and dS represents the boundary of S, the entire object can be represented as a collection of its boundaries dS. The entities that constitute dS can be points, edges, or surfaces homologous to R 0 , R 1 , or R 2 , in the Euclidean space [7].…”