SplatReasoner: Enhancing Embodied Reasoning and Grounding by Novel View Synthesis

1POSTECH 2KAIST 3ETRI 4NVIDIA
ECCV 2026
Overview of the SplatReasoner framework

SplatReasoner aims to enhance embodied reasoning and 3D visual grounding by utilizing novel-view synthesis to improve the 3D scene reasoning capabilities of Vision-Language Models (VLMs).

Abstract

Vision-Language Models (VLMs) have demonstrated strong reasoning capabilities over images and videos, yet their application to embodied scene understanding is often constrained by the fixed viewpoints stored in episodic RGB-D memories. These observations may fail to capture query-relevant evidence due to occlusions, object truncation, restricted fields of view, or suboptimal view composition. We present SplatReasoner, a framework that introduces novel view synthesis into the VLM reasoning process by leveraging 3D Gaussian Splatting (3DGS). Given a user query about a 3D scene, SplatReasoner retrieves relevant observations and synthesizes query-conditioned viewpoints that reveal the visual evidence needed to answer the query and ground the referred entities in 3D. Experiments show that query-conditioned novel view synthesis improves both embodied reasoning and 3D grounding over fixed-view memory and language-embedded 3DGS baselines.

Overview

Overview of the SplatReasoner view selection pipeline

Given \(N\) pre-captured views, we sample \(L'\) images to perform embodied tasks. First, we select \(L\) initial views from \(N\) input views through initial view selection (\(N \gg L\)). Then, we augment each initial view with \(V\) novel views. Finally, we proceed with the final view selection with VLMs to find the most informative views to answer the query.

Initial View Selection in Language-embedded 3DGS

Initial view selection with language-embedded 3D Gaussian Splatting

First, we identify the activated Gaussians \(\mathcal{G}^{\mathrm{act}}\) using a set of evidence categories \(\mathcal{C}^{\mathrm{evidence}}\) extracted from the query \(\mathcal{Q}\). Next, these activated Gaussians are grouped into clusters \(\{\mathcal{G}^{\mathrm{cluster}}_l\}_{l=1}^{L}\) that serve as instance-level candidates. Finally, the initial viewpoints are determined by maximizing the visibility score.

Final View Selection with Novel View Synthesis

Final view selection with novel view synthesis

For each initial view, we synthesize novel views and employ a VLM-as-Judge to select the most informative viewpoints from both the initial and novel view candidates. To ensure that the view selected by the VLM-as-Judge provides richer information than the initial one, we conduct a verification stage.

EM-EQA: Episodic Memory Embodied Question Answering

LLM-Match Higher is better
Closed Model (GPT-4o) Open Model (Qwen3-VL 8B)
BlindLLM
34.8
29.4
0frame
Frame Captions
24.1
31.7
0frame
CG Captions*
36.5
0frame
SVM Captions*
38.9
0frame
Multi-Frame
49.1
48.2
3.0frames
3D-Mem
54.6
50.8
2.7frames
SplatReasoner (ours)
57.8
51.6
2.6frames
Human
86.8
Full

Methods marked with * and the Human score are reported from OpenEQA, while all other results are reproduced. We report OpenEQA results since the code is not publicly available. Average frames denote the number of final views used to answer queries. indicates that while novel views are utilized for the VLM-as-Judge, the actual number of input frames fed into the VLMs remains consistent with the reported average frames.

3D Visual Grounding

Qualitative 3D visual grounding results on ScanRefer
ScanRefer Bars scaled per metric · Higher is better
Method 3D mIoU Acc@8 Acc@10
Dr. Splat (category)
8.73
34.53
28.53
Dr. Splat (sentence)
9.44
34.28
29.35
ReferSplat
3.14
10.68
7.32
SplatReasoner ours
11.12
35.84
32.75

Acc@k denotes the percentage of predictions whose IoU with the ground truth exceeds k%.

BibTeX


    @inproceedings{yu2026splatreasoner,
        title={SplatReasoner: Enhancing Embodied Reasoning and Grounding by Novel View Synthesis},
        author={Yu-Ji, Kim and Lee, Dahye and Jun-Seong, Kim and Hyeon-Woo, Nam and Kim, GeonU and Kwon, Yongjin and Wang, Yu-Chiang Frank and Choe, Jaesung and Oh, Tae-Hyun},
        booktitle={ECCV},
        year={2026}
    }