PointNSP: Autoregressive 3D Point Cloud Generation with Next-Scale Level-of-Detail Prediction
This addresses the challenge of generating high-quality 3D point clouds efficiently for applications like computer graphics and robotics, representing a significant advance rather than an incremental improvement.
The paper tackled the problem of autoregressive point cloud generation lagging behind diffusion-based methods by proposing PointNSP, a coarse-to-fine framework that uses next-scale prediction to preserve global structure and refine geometry, achieving state-of-the-art quality on ShapeNet and surpassing diffusion baselines in efficiency, with advantages in dense generation up to 8,192 points.
Autoregressive point cloud generation has long lagged behind diffusion-based approaches in quality. The performance gap stems from the fact that autoregressive models impose an artificial ordering on inherently unordered point sets, forcing shape generation to proceed as a sequence of local predictions. This sequential bias emphasizes short-range continuity but undermines the model's capacity to capture long-range dependencies, hindering its ability to enforce global structural properties such as symmetry, consistent topology, and large-scale geometric regularities. Inspired by the level-of-detail (LOD) principle in shape modeling, we propose PointNSP, a coarse-to-fine generative framework that preserves global shape structure at low resolutions and progressively refines fine-grained geometry at higher scales through a next-scale prediction paradigm. This multi-scale factorization aligns the autoregressive objective with the permutation-invariant nature of point sets, enabling rich intra-scale interactions while avoiding brittle fixed orderings. Experiments on ShapeNet show that PointNSP establishes state-of-the-art (SOTA) generation quality for the first time within the autoregressive paradigm. In addition, it surpasses strong diffusion-based baselines in parameter, training, and inference efficiency. Finally, in dense generation with 8,192 points, PointNSP's advantages become even more pronounced, underscoring its scalability potential.