Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Gong, Ziren, Li, Xiaohan, Tosi, Fabio, Han, Jiawei, Mattoccia, Stefano, Cai, Jianfei, Poggi, Matteo
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2507.22052
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911297467056128
author Gong, Ziren
Li, Xiaohan
Tosi, Fabio
Han, Jiawei
Mattoccia, Stefano
Cai, Jianfei
Poggi, Matteo
author_facet Gong, Ziren
Li, Xiaohan
Tosi, Fabio
Han, Jiawei
Mattoccia, Stefano
Cai, Jianfei
Poggi, Matteo
contents We present Ov3R, a novel framework for open-vocabulary semantic 3D reconstruction from RGB video streams, designed to advance Spatial AI. The system features two key components: CLIP3R, a CLIP-informed 3D reconstruction module that predicts dense point maps from overlapping clips while embedding object-level semantics; and 2D-3D OVS, a 2D-3D open-vocabulary semantic module that lifts 2D features into 3D by learning fused descriptors integrating spatial, geometric, and semantic cues. Unlike prior methods, Ov3R incorporates CLIP semantics directly into the reconstruction process, enabling globally consistent geometry and fine-grained semantic alignment. Our framework achieves state-of-the-art performance in both dense 3D reconstruction and open-vocabulary 3D segmentation, marking a step forward toward real-time, semantics-aware Spatial AI.
format Preprint
id arxiv_https___arxiv_org_abs_2507_22052
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Ov3R: Open-Vocabulary Semantic 3D Reconstruction from RGB Videos
Gong, Ziren
Li, Xiaohan
Tosi, Fabio
Han, Jiawei
Mattoccia, Stefano
Cai, Jianfei
Poggi, Matteo
Computer Vision and Pattern Recognition
We present Ov3R, a novel framework for open-vocabulary semantic 3D reconstruction from RGB video streams, designed to advance Spatial AI. The system features two key components: CLIP3R, a CLIP-informed 3D reconstruction module that predicts dense point maps from overlapping clips while embedding object-level semantics; and 2D-3D OVS, a 2D-3D open-vocabulary semantic module that lifts 2D features into 3D by learning fused descriptors integrating spatial, geometric, and semantic cues. Unlike prior methods, Ov3R incorporates CLIP semantics directly into the reconstruction process, enabling globally consistent geometry and fine-grained semantic alignment. Our framework achieves state-of-the-art performance in both dense 3D reconstruction and open-vocabulary 3D segmentation, marking a step forward toward real-time, semantics-aware Spatial AI.
title Ov3R: Open-Vocabulary Semantic 3D Reconstruction from RGB Videos
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.22052