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

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Monchieri, Leonardo, Camuffo, Elena, Barbato, Francesco, Zanuttigh, Pietro, Milani, Simone
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908809890365440
author Monchieri, Leonardo
Camuffo, Elena
Barbato, Francesco
Zanuttigh, Pietro
Milani, Simone
author_facet Monchieri, Leonardo
Camuffo, Elena
Barbato, Francesco
Zanuttigh, Pietro
Milani, Simone
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2602_03809
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Split&Splat: Zero-Shot Panoptic Segmentation via Explicit Instance Modeling and 3D Gaussian Splatting
Monchieri, Leonardo
Camuffo, Elena
Barbato, Francesco
Zanuttigh, Pietro
Milani, Simone
Graphics
Computer Vision and Pattern Recognition
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.
title Split&Splat: Zero-Shot Panoptic Segmentation via Explicit Instance Modeling and 3D Gaussian Splatting
topic Graphics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.03809