Variable Rate Compression for Raw 3D Point Clouds
Md Ahmed Al Muzaddid,William J. Beksi,Md Ahmed Al Muzaddid,William J. Beksi
In this paper, we propose a novel variable rate deep compression architecture that operates on raw 3D point cloud data. The majority of learning-based point cloud compression methods work on a downsampled representation of the data. Moreover, many existing techniques require training multiple networks for different compression rates to generate consolidated point clouds of varying quality. In cont...
