{"ID":2859173,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.05613","arxiv_id":"2510.05613","title":"PointNSP: Autoregressive 3D Point Cloud Generation with Next-Scale Level-of-Detail Prediction","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.","short_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...","url_abs":"https://arxiv.org/abs/2510.05613","url_pdf":"https://arxiv.org/pdf/2510.05613v2","authors":"[\"Ziqiao Meng\",\"Qichao Wang\",\"Zhiyang Dou\",\"Zixing Song\",\"Zhipeng Zhou\",\"Irwin King\",\"Peilin Zhao\"]","published":"2025-10-07T06:31:02Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\",\"cs.LG\"]","methods":"[\"Diffusion Model\"]","has_code":false}