RT-GS2: Real-Time Generalizable Semantic Segmentation for 3D Gaussian Representations of Radiance Fields

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jurca, Mihnea-Bogdan, Royen, Remco, Giosan, Ion, Munteanu, Adrian
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916375213113344
author Jurca, Mihnea-Bogdan
Royen, Remco
Giosan, Ion
Munteanu, Adrian
author_facet Jurca, Mihnea-Bogdan
Royen, Remco
Giosan, Ion
Munteanu, Adrian
contents Gaussian Splatting has revolutionized the world of novel view synthesis by achieving high rendering performance in real-time. Recently, studies have focused on enriching these 3D representations with semantic information for downstream tasks. In this paper, we introduce RT-GS2, the first generalizable semantic segmentation method employing Gaussian Splatting. While existing Gaussian Splatting-based approaches rely on scene-specific training, RT-GS2 demonstrates the ability to generalize to unseen scenes. Our method adopts a new approach by first extracting view-independent 3D Gaussian features in a self-supervised manner, followed by a novel View-Dependent / View-Independent (VDVI) feature fusion to enhance semantic consistency over different views. Extensive experimentation on three different datasets showcases RT-GS2's superiority over the state-of-the-art methods in semantic segmentation quality, exemplified by a 8.01% increase in mIoU on the Replica dataset. Moreover, our method achieves real-time performance of 27.03 FPS, marking an astonishing 901 times speedup compared to existing approaches. This work represents a significant advancement in the field by introducing, to the best of our knowledge, the first real-time generalizable semantic segmentation method for 3D Gaussian representations of radiance fields.
format Preprint
id arxiv_https___arxiv_org_abs_2405_18033
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle RT-GS2: Real-Time Generalizable Semantic Segmentation for 3D Gaussian Representations of Radiance Fields
Jurca, Mihnea-Bogdan
Royen, Remco
Giosan, Ion
Munteanu, Adrian
Computer Vision and Pattern Recognition
Gaussian Splatting has revolutionized the world of novel view synthesis by achieving high rendering performance in real-time. Recently, studies have focused on enriching these 3D representations with semantic information for downstream tasks. In this paper, we introduce RT-GS2, the first generalizable semantic segmentation method employing Gaussian Splatting. While existing Gaussian Splatting-based approaches rely on scene-specific training, RT-GS2 demonstrates the ability to generalize to unseen scenes. Our method adopts a new approach by first extracting view-independent 3D Gaussian features in a self-supervised manner, followed by a novel View-Dependent / View-Independent (VDVI) feature fusion to enhance semantic consistency over different views. Extensive experimentation on three different datasets showcases RT-GS2's superiority over the state-of-the-art methods in semantic segmentation quality, exemplified by a 8.01% increase in mIoU on the Replica dataset. Moreover, our method achieves real-time performance of 27.03 FPS, marking an astonishing 901 times speedup compared to existing approaches. This work represents a significant advancement in the field by introducing, to the best of our knowledge, the first real-time generalizable semantic segmentation method for 3D Gaussian representations of radiance fields.
title RT-GS2: Real-Time Generalizable Semantic Segmentation for 3D Gaussian Representations of Radiance Fields
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2405.18033