English

PointNSP: Autoregressive 3D Point Cloud Generation with Next-Scale Level-of-Detail Prediction

Computer Vision and Pattern Recognition 2025-12-02 v3 Machine Learning

Abstract

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.

Keywords

Cite

@article{arxiv.2503.08594,
  title  = {PointNSP: Autoregressive 3D Point Cloud Generation with Next-Scale Level-of-Detail Prediction},
  author = {Ziqiao Meng and Qichao Wang and Zhiyang Dou and Zixing Song and Zhipeng Zhou and Irwin King and Peilin Zhao},
  journal= {arXiv preprint arXiv:2503.08594},
  year   = {2025}
}

Comments

24 pages; Previously this version appeared as arXiv:2510.05613 which was submitted as a new work by accident