Dynamic compression of curve-based point cloud

摘要: With the increasing demands for highly detailed 3D data, dynamic scanning systems are capable of producing 3D+t (a.k.a. 4D) spatio-temporal models with millions points recently. As a consequence, effective 4D geometry compression schemes required to face need store/transmit huge amount in addition classical static data. In this paper, we propose point cloud encoder via curve-based representation cloud, particularly well-suited structured-light-based systems, wherein grid pattern is projected onto surface object. The object then naturally sampled series curves, due pattern. This motivates our choice leverage remove spatial and temporal correlation along directions through competitive-based predictive that includes different prediction modes. Experimental results show significant gain obtained proposed method.