Density-preserving Deep Point Cloud Compression

TL;DR

This paper introduces a deep point cloud compression method that preserves local density information, improving the quality of compressed point clouds by explicitly encoding local geometry and density, and employing novel upsampling techniques.

Contribution

The proposed method uniquely encodes local density and geometry with three embeddings and introduces a sub-point convolution layer for better upsampling, achieving state-of-the-art results.

Findings

Achieves superior rate-distortion trade-off on SemanticKITTI and ShapeNet datasets.

Effectively preserves local density and geometry in compressed point clouds.

Outperforms existing methods in both qualitative and quantitative evaluations.

Abstract

Local density of point clouds is crucial for representing local details, but has been overlooked by existing point cloud compression methods. To address this, we propose a novel deep point cloud compression method that preserves local density information. Our method works in an auto-encoder fashion: the encoder downsamples the points and learns point-wise features, while the decoder upsamples the points using these features. Specifically, we propose to encode local geometry and density with three embeddings: density embedding, local position embedding and ancestor embedding. During the decoding, we explicitly predict the upsampling factor for each point, and the directions and scales of the upsampled points. To mitigate the clustered points issue in existing methods, we design a novel sub-point convolution layer, and an upsampling block with adaptive scale. Furthermore, our method can also compress point-wise attributes, such as normal. Extensive qualitative and quantitative results on SemanticKITTI and ShapeNet demonstrate that our method achieves the state-of-the-art rate-distortion trade-off.