DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processing
arXiv:1909.03669
Abstract
Point cloud processing is very challenging, as the diverse shapes formed by irregular points are often indistinguishable. A thorough grasp of the elusive shape requires sufficiently contextual semantic information, yet few works devote to this. Here we propose DensePoint, a general architecture to learn densely contextual representation for point cloud processing. Technically, it extends regular grid CNN to irregular point configuration by generalizing a convolution operator, which holds the permutation invariance of points, and achieves efficient inductive learning of local patterns. Architecturally, it finds inspiration from dense connection mode, to repeatedly aggregate multi-level and multi-scale semantics in a deep hierarchy. As a result, densely contextual information along with rich semantics, can be acquired by DensePoint in an organic manner, making it highly effective. Extensive experiments on challenging benchmarks across four tasks, as well as thorough model analysis, verify DensePoint achieves the state of the arts.
arXiv commentAccepted to ICCV 2019. 15 pages, 8 figures, 16 tables
References in corpus (6)
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- PCPNET: Learning Local Shape Properties from Raw Point Clouds
- PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation
- Deep Roots: Improving CNN Efficiency with Hierarchical Filter Groups
- SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation
- Deep Learning with Sets and Point Clouds