ViewFusion: Structured Spatial Thinking Chains for Multi-View Reasoning

Xingjian Tao; Yiwei Wang; Yujun Cai; Yifan Song; Jing Tang

cs.CLMar 6, 2026

ViewFusion: Structured Spatial Thinking Chains for Multi-View Reasoning

Authors: Xingjian Tao, Yiwei Wang, Yujun Cai, Yifan Song, Jing Tang

Organizations: HKUST(GZ) · UC Merced · University of Queensland · HKUST

Abstract

Multi-view spatial reasoning remains difficult for current vision-language models. Even when multiple viewpoints are available, models often underutilize cross-view relations and instead rely on single-image shortcuts, leading to fragile performance on viewpoint transformation and occlusion-sensitive cases. We present ViewFusion, a two-stage framework that explicitly separates cross-view spatial pre-alignment from question answering. In the first stage, the model performs deliberate spatial pre-thinking to infer viewpoint relations and spatial transformations across views, forming an intermediate workspace that goes beyond a simple re-description. In the second stage, the model conducts question-driven reasoning conditioned on this workspace to produce the final prediction. We train ViewFusion with synthetic reasoning supervision followed by reinforcement learning using GRPO, which improves answer correctness while stabilizing the intended two-stage generation behavior. On MMSI-Bench, ViewFusion improves accuracy by 5.3% over Qwen3-VL-4B-Instruct, with the largest gains on examples that require genuine cross-view alignment.

Figures & tables

Figure 1: An example of multi-view spatial reasoning. Given two images captured from different viewpoints, the model must align shared visual cues across views to infer the viewpoint relationship and answer a direction-based question (e.g., locating the picture frame relative to the piano)

Figure 2: Overview of ViewFusion for multi-view spatial reasoning. Given a multi-view question (left), existing “describe-first” or direct “think-and-answer” paradigms often produce view-local descriptions and then shortcut to answering without establishing correct cross-view spatial relations, leading to errors (top). ViewFusion instead performs explicit multi-view spatial pre-thinking to link perspectives and infer viewpoint transformations across images before question solving (bottom), yielding more reliable reasoning and correct predictions.

Close-source models
Model	Size	MMSI	MindCube	ViewSpatial	Avg.
RandomChoice	–	25.0	33.0	26.3	28.1
Gemini-2.5-Pro	–	38.0	57.6	46.0	47.2
GPT-5	–	41.8	56.3	45.5	47.9
Gemini-3-Pro-Preview	–	45.2	70.8	50.3	55.4
Open-source General Models

Table 1: Overall accuracy (%) on three multi-view spatial reasoning benchmarks (MMSI-Bench, MindCube, and ViewSpatial) for a range of proprietary and open-source MLLMs, including our ViewFusion variants

Figure 3: Qualitative examples on MMSI-Bench. The red boxes highlight the same visual elements observed from different viewpoints across the two images. Compared with Qwen3-VL-4B-Instruct, ViewFusion better aligns cross-view correspondences and infers the underlying viewpoint change, leading to correct answers.

Models	Positional Relationship						Attribute		Motion		MSR	Avg.
Models	Cam.–Cam.	Obj.–Obj.	Reg.–Reg.	Cam.–Obj.	Obj.–Reg.	Cam.–Reg.	Meas.	Appr.	Cam.	Obj.	–
Qwen3-4B-VL-Thinking	25.8	26.6	34.5	33.7	25.9	36.1	48.4	28.8	21.6	26.3	23.2	29.0
Qwen3-4B-VL-Instruct	30.1	34.0	29.6	34.9	29.4	39.8	45.3	19.7	21.6	23.7	26.7	30.1
ViewFusion	46.2	41.5	30.9	44.2	21.2	53.0	35.9	34.9	40.5	32.9	23.2	35.4

Table 2: Fine-grained accuracy (%) breakdown on MMSI-Bench, comparing ViewFusion with Qwen3-VL-4B-Instruct and Qwen3-VL-4B-Thinking across subcategories of positional relationships, attributes, motion, and MSR.

Models	Positional Relationship						Attribute		Motion		MSR	Avg.
Models	Cam.–Cam.	Obj.–Obj.	Reg.–Reg.	Cam.–Obj.	Obj.–Reg.	Cam.–Reg.	Meas.	Appr.	Cam.	Obj.	–
Full Method	46.2	41.5	30.9	44.2	21.2	53.0	35.9	34.9	40.5	32.9	23.2	35.4
Free format Reasoning + RL	37.6	31.9	25.9	39.5	28.2	49.4	46.9	27.3	31.1	29.0	28.3	33.4
w/o GRPO	28.0	27.7	35.8	37.2	29.4	47.0	45.3	30.3	32.4	25.0	27.8	32.4
w/o Format Reward	40.9	30.9	33.3	46.5	32.9	51.8	29.7	40.9	37.8	34.2	22.7	35.0

Table 3: Ablation study on MMSI-Bench (accuracy, %), analyzing the impact of free-form reasoning, removing GRPO, and removing the format reward on fine-grained subcategories and overall performance.

Panel of Figure 4 — Figure 4: Training curves during GRPO over 1500 steps, including the total reward (left), the accuracy reward (second), the format reward (third), and the KL divergence to the reference policy (right).

Visual Spatial Reasoning

Visual spatial reasoning studies how AI systems infer geometry, object relations, viewpoints, and physical change from images and video.

See all 251 papers

Seek-and-View Reasoning for Multi-View Spatial Understanding
Oct 8, 2026Qixiang Chen, Cheng Zhang, Fucai Ke +3Multimodal Spatial Reasoning Visual Spatial Reasoning
Monash University · The Chinese University of Hong Kong
Learning Multi-View Spatial Reasoning from Cross-View Relations
Mar 30, 2026Suchae Jeong, Jaehwi Song, Haeone Lee +9Multimodal Spatial Reasoning Visual Spatial Reasoning
KAIST · Config · Hanyang University +2
CrossView Suite: Harnessing Cross-view Spatial Intelligence of MLLMs with Dataset, Model and Benchmark
May 18, 2026Wei Wang, Yuqian Yuan, Tianwei Lin +4VLM Evaluation Multi-View Consistency
Zhejiang University China

Oct 8, 2026J/K move · Enter open · S save

Seek-and-View Reasoning for Multi-View Spatial Understanding

Qixiang Chen, Cheng Zhang, Fucai Ke, Chi-Wing Fu+2

Monash University · The Chinese University of Hong Kong

Multimodal Spatial Reasoning Visual Spatial Reasoning Spatial Reasoning VLM Reasoning

Open paper AskPDF arXiv

Existing approaches to multi-view spatial reasoning operate largely on sparse input views. Vision-language models (VLMs) are thus restricted to understand a scene and infer spatial relations within these fixed views, leading to fragile cross-view alignment and geometry-to-language bottleneck. To address these issues, we formulate a novel Seek-and-View reasoning approach to find implicit cross-view spatial evidence by locating a question-relevant view to support the spatial reasoning. To realize this approach, we propose Vantage, a training-free model-agnostic reasoning framework that pairs a VLM with a 3D foundation model: a viewpoint-grounded reasoning stage for question analysis and view planning, followed by a geometry-grounded evidence augmentation stage to effectively synthesize and incorporate visual evidence into the final reasoning. Comprehensive experiments on six VLMs demonstrate consistent improvements on five benchmarks without fine-tuning. Overall, by revealing spatial evidence through view-grounded reasoning, Vantage can largely reduce reliance on language-based cross-view alignment and improve multi-view spatial understanding. Our code is available at https://github.com/q1xiangchen/Vantage.

Abstract

Visual Spatial Reasoning

Explore similar work

Seek-and-View Reasoning for Multi-View Spatial Understanding

Learning Multi-View Spatial Reasoning from Cross-View Relations

CrossView Suite: Harnessing Cross-view Spatial Intelligence of MLLMs with Dataset, Model and Benchmark