Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jun-Seong, Kim, Kim, GeonU, Yu-Ji, Kim, Wang, Yu-Chiang Frank, Choe, Jaesung, Oh, Tae-Hyun
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916626179293184
author Jun-Seong, Kim
Kim, GeonU
Yu-Ji, Kim
Wang, Yu-Chiang Frank
Choe, Jaesung
Oh, Tae-Hyun
author_facet Jun-Seong, Kim
Kim, GeonU
Yu-Ji, Kim
Wang, Yu-Chiang Frank
Choe, Jaesung
Oh, Tae-Hyun
contents We introduce Dr. Splat, a novel approach for open-vocabulary 3D scene understanding leveraging 3D Gaussian Splatting. Unlike existing language-embedded 3DGS methods, which rely on a rendering process, our method directly associates language-aligned CLIP embeddings with 3D Gaussians for holistic 3D scene understanding. The key of our method is a language feature registration technique where CLIP embeddings are assigned to the dominant Gaussians intersected by each pixel-ray. Moreover, we integrate Product Quantization (PQ) trained on general large-scale image data to compactly represent embeddings without per-scene optimization. Experiments demonstrate that our approach significantly outperforms existing approaches in 3D perception benchmarks, such as open-vocabulary 3D semantic segmentation, 3D object localization, and 3D object selection tasks. For video results, please visit : https://drsplat.github.io/
format Preprint
id arxiv_https___arxiv_org_abs_2502_16652
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration
Jun-Seong, Kim
Kim, GeonU
Yu-Ji, Kim
Wang, Yu-Chiang Frank
Choe, Jaesung
Oh, Tae-Hyun
Computer Vision and Pattern Recognition
We introduce Dr. Splat, a novel approach for open-vocabulary 3D scene understanding leveraging 3D Gaussian Splatting. Unlike existing language-embedded 3DGS methods, which rely on a rendering process, our method directly associates language-aligned CLIP embeddings with 3D Gaussians for holistic 3D scene understanding. The key of our method is a language feature registration technique where CLIP embeddings are assigned to the dominant Gaussians intersected by each pixel-ray. Moreover, we integrate Product Quantization (PQ) trained on general large-scale image data to compactly represent embeddings without per-scene optimization. Experiments demonstrate that our approach significantly outperforms existing approaches in 3D perception benchmarks, such as open-vocabulary 3D semantic segmentation, 3D object localization, and 3D object selection tasks. For video results, please visit : https://drsplat.github.io/
title Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.16652