Split&Splat: Zero-Shot Panoptic Segmentation via Explicit Instance Modeling and 3D Gaussian Splatting

Leonardo Monchieri, Elena Camuffo, Francesco Barbato, Pietro Zanuttigh, Simone MilaniarXiv 2026 · arXiv

Abstract

3D Gaussian Splatting (GS) enables fast and high-quality scene reconstruction, but it lacks an object-consistent and semantically aware structure. We propose Split&Splat, a framework for panoptic scene reconstruction using 3DGS. Our approach explicitly models object instances. It first propagates instance masks across views using depth, thus producing view-consistent 2D masks. Each object is then reconstructed independently and merged back into the scene while refining its boundaries. Finally, instance-level semantic descriptors are embedded in the reconstructed objects, supporting various applications, including panoptic segmentation, object retrieval, and 3D editing. Unlike existing methods, Split&Splat tackles the problem by first segmenting the scene and then reconstructing each object individually. This design naturally supports downstream tasks and allows Split&Splat to achieve state-of-the-art performance on the ScanNetv2 segmentation benchmark.
BibTeX
@misc{monchieri2026split,
  author = {Leonardo Monchieri and Elena Camuffo and Francesco Barbato and Pietro Zanuttigh and Simone Milani},
  title = {{Split\&Splat: Zero-Shot Panoptic Segmentation via Explicit Instance Modeling and 3D Gaussian Splatting}},
  year = {2026},
  eprint = {2602.03809},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2602.03809},
}

Loading related content...

Figure