Precision Controllable Point Clouds Compression Using Global Similarity in Dictionary

To control the precision of the point clouds compression, this paper proposes an algorithm for precision controllable point clouds compression using the global similarity in dictionary, which is built on the uniform spatial partitioning of the point cloud. Firstly, the data is discretely represented by uniform spatial partitioning. Secondly, a dictionary is created as the bases of the discrete data, upon which a data structure can be built using the indices. Finally, the dictionary is compressed by the combination of the vocabulary entries, and the indices are compressed by a lossless encoding algorithm. Experimental results show that our method performs compression with arbitrarily specified precision on two sets of data from ModelNet and Farman datasets under a common personal computer environment.