OctAttention: Octree-Based Large-Scale Contexts Model for Point Cloud Compression
This work addresses compression efficiency for sparse point clouds, which is crucial for applications like LiDAR and 3D object representation, though it appears incremental as it builds on existing octree and attention mechanisms.
The paper tackles the problem of insufficient context modeling in point cloud compression for sparse point clouds by proposing OctAttention, an octree-based multiple-contexts deep learning framework. The result is a 10%-35% BD-Rate gain on benchmarks like SemanticKITTI and MPEG 8i, and a 95% reduction in coding time compared to voxel-based methods.
In point cloud compression, sufficient contexts are significant for modeling the point cloud distribution. However, the contexts gathered by the previous voxel-based methods decrease when handling sparse point clouds. To address this problem, we propose a multiple-contexts deep learning framework called OctAttention employing the octree structure, a memory-efficient representation for point clouds. Our approach encodes octree symbol sequences in a lossless way by gathering the information of sibling and ancestor nodes. Expressly, we first represent point clouds with octree to reduce spatial redundancy, which is robust for point clouds with different resolutions. We then design a conditional entropy model with a large receptive field that models the sibling and ancestor contexts to exploit the strong dependency among the neighboring nodes and employ an attention mechanism to emphasize the correlated nodes in the context. Furthermore, we introduce a mask operation during training and testing to make a trade-off between encoding time and performance. Compared to the previous state-of-the-art works, our approach obtains a 10%-35% BD-Rate gain on the LiDAR benchmark (e.g. SemanticKITTI) and object point cloud dataset (e.g. MPEG 8i, MVUB), and saves 95% coding time compared to the voxel-based baseline. The code is available at https://github.com/zb12138/OctAttention.