FlowScene: Style-Consistent Indoor Scene Generation with Multimodal Graph Rectified Flow

Zhifei Yang; Guangyao Zhai; Keyang Lu; YuYang Yin; Chao Zhang; Zhen Xiao; Jieyi Long; Nassir Navab

cs.CVMar 20, 2026

FlowScene: Style-Consistent Indoor Scene Generation with Multimodal Graph Rectified Flow

Authors: Zhifei Yang, Guangyao Zhai, Keyang Lu, YuYang Yin, Chao Zhang, Zhen Xiao, Jieyi Long, Nassir Navab,

Organizations: School of Computer Science, Peking University · Technical University of Munich · Beijing Jiaotong University · Beijing Digital Native Digital City Research Center ·

Abstract

Scene generation has extensive industrial applications, demanding both high realism and precise control over geometry and appearance. Language-driven retrieval methods compose plausible scenes from a large object database, but overlook object-level control and often fail to enforce scene-level style coherence. Graph-based formulations offer higher controllability over objects and inform holistic consistency by explicitly modeling relations, yet existing methods struggle to produce high-fidelity textured results, thereby limiting their practical utility. We present FlowScene, a tri-branch scene generative model conditioned on multimodal graphs that collaboratively generates scene layouts, object shapes, and object textures. At its core lies a tight-coupled rectified flow model that exchanges object information during generation, enabling collaborative reasoning across the graph. This enables fine-grained control of objects' shapes, textures, and relations while enforcing scene-level style coherence across structure and appearance. Extensive experiments show that FlowScene outperforms both language-conditioned and graph-conditioned baselines in terms of generation realism, style consistency, and alignment with human preferences.

Figures & tables

Figure 1. Scene Generation from Diverse Input. The prospective system, powered by FlowScene , supports the generation of style-consistent 3D scenes from multi-source descriptions, including text input, GUI selections, and mixed information. Users can flexibly specify object categories and, if desired, provide their detailed RGB images in arbitrary views, along with their optional inter-object relationships (A). These descriptions are parsed into a multimodal graph by modern LLMs or VLMs as graph constructors (B). Based on this graph, FlowScene generates complete textured scenes, comprising both spatial scene layout and individual posed objects (C).

Figure 2. Multimodal Graph Processing. Given a graph $\mathcal{G}_{\mathrm{M}}$ (upper left), the graph network (right) first projects multimodal features of nodes $\mathcal{V}_{\mathrm{M}}$ , followed by message passing and feature aggregation along the edges $\mathcal{E}$ connecting $\mathcal{V}_{\mathrm{M}}$ . Within the denoising process of a generative model, it acts as an InfoExchangeUnit to incorporate temporal denoising data $\mathcal{D}_{t}$ , which is projected and fused with the node features (lower left). In this way, the network yields node-wise conditioning $\mathcal{C}_{t}$ while respecting the global constraints imposed by the graph structure (bottom).

Figure 3. FlowScene Pipeline. Conditioned on $\mathcal{G}_{\mathrm{M}}$ in Figure 2 , FlowScene stacks three coordinated branches to generate a scene. A. Layout Branch generates the scene layout from noise $\mathcal{B}_{1}$ , where the x-axis (red) points to the right and the z-axis (blue) points to the front. The LayoutExchangeUnit iteratively applies temporal layout constraints to the generation process by conditioning $\Theta_{\mathcal{B}}$ . B. Shape Branch is parallel to the layout branch. It exchanges shape information among nodes according to $\mathcal{G}_{\mathrm{M}}$ through the ShapeExchangeUnit . At inference, we sample noise $\mathcal{S}_{1}$ , and generate latent shape codes $\mathcal{S}$ via $\Theta_{\mathcal{S}}$ , and decode them into voxelized shapes $\mathcal{X}$ using $\Phi_{D}$ . C. Texture Branch is subsequent to the shape branch. It anchors Gaussian noise onto $\mathcal{X}$ to form $\mathcal{X}_{1}^{e}$ , which is then denoised by $\Theta_{\mathcal{T}}$ into structured latents $\mathcal{X}^{e}$ . Finally, $\Psi_{D}$ decodes them into objects $\mathcal{O}$ , which are scaled and placed into the generated layout $\mathcal{B}$ to complete the scene.

Method	Bedroom			Living room			Dining room
Method	FID	$\text{FID}_{\text{CLIP}}$	KID	FID	$\text{FID}_{\text{CLIP}}$	KID	FID	$\text{FID}_{\text{CLIP}}$	KID
CommonScenes ∗ ( Zhai et al., 2024c )	49.52	0 5.56	0 2.20	78.19	0 7.59	0 6.12	69.89	0 7.88	0 4.78
EchoScene ∗ ( Zhai et al., 2024b )	44.10	0 4.49	0 0.30	75.35	0 7.37	0 5.59	67.52	0 7.27	0 3.15
MMGDreamer ∗ ( Yang et al., 2025a )	42.38	0 4.48	0 -0.14	74.14	0 7.31	0 5.62	67.97	0 7.39	0 2.93
FlowScene (Ours)	35.01	0 3.14	0 -0.34	70.34	0 7.18	0 5.11	60.26	0 6.72	0 0.07
CommonScenes ( Zhai et al., 2024c )	57.68	0 4.86	0 6.59	80.99	0 7.05	0 6.39	65.71	0 7.04	0 5.47

Table 1. Scene-Level Fidelity. This is evaluated by comparing top-down renderings of generated and real scenes at

256^{2}

resolution using FID,

\text{FID}_{\text{CLIP}}

, and KID (

\times 0.001

), following ( Zhai et al., 2024b ) (lower is better). * denotes their retrieval branches, which generate layouts and retrieve well-designed meshes from 3D-FUTURE ( Fu et al., 2021b ) . The best results are in bold .

Method	Metric	Bed	N.stand	Ward.	Chair	Table	Cabinet	Lamp	Shelf	Sofa	TV stand
CommonScenes ( Zhai et al., 2024c )	MMD ( $\downarrow$ )	0.49	0.92	0.54	0.99	1.91	0.96	1.50	2.73	0.57	0.29
EchoScene ( Zhai et al., 2024b )		0.37	0.75	0.39	0.62	1.47	0.83	0.66	2.52	0.48	0.35
MMGDreamer ( Yang et al., 2025a )		0.22	0.41	0.24	0.35	0.55	0.71	0.34	1.58	0.43	0.24
FlowScene (Ours)		0.17	0.23	0.20	0.32	0.64	0.60	0.20	1.53	0.48	0.13
CommonScenes ( Zhai et al., 2024c )	COV ( $\%,\uparrow$ )	24.07	24.17	26.62	26.72	40.52	28.45	36.21	40.00	28.45	33.62
EchoScene ( Zhai et al., 2024b )		39.51	25.59	37.07	17.25	35.05	43.21	33.33	50.00	41.94	40.70

Table 2. Object-Level Fidelity. Quantitative evaluation of generated object quality and diversity using MMD (

\times 0.01

), COV, and 1-NNA metrics across different categories. The best results are highlighted in bold .

Method	CLIPScore $\uparrow$			FPVScore $\uparrow$					Perceptual Study $\uparrow$
	Bedroom	Living room	Dining room	PA	LC	VQ	SC	OP	PA	LC	VQ	SC	OP
Echoscene ∗ ( Zhai et al., 2024b )	0.1837	0.1799	0.1553	3.03	3.07	2.82	3.03	3.01	7.21	6.83	6.86	7.12	7.24
MMGDreamer ∗ ( Yang et al., 2025a )	0.1957	0.1877	0.1626	3.12	3.31	3.08	3.25	3.27	7.47	7.12	7.25	7.31	7.42
Holodeck ( Yang et al., 2024c )	0.1912	0.2116	0.1968	2.79	2.73	2.71	2.70	2.70	7.16	6.61	7.43	6.72	6.64
LayoutVLM ( Sun et al., 2025a )	0.2045	0.2056	0.1947	2.36	2.44	2.15	2.23	2.31	6.52	6.33	7.21	6.43	6.32
FlowScene (Ours)	0.2386	0.2246	0.2092	3.35	3.54	3.83	3.85	3.76	7.97	7.65	8.46	8.72	8.57

Table 3. Scene Perception Evaluation. For FPVScore, we report the average ranking across all scenes, where 1.00 is the worst and 5.00 is the best. For the perceptual study, scores range from 1.00 to 10.00. Each score is averaged across all participants and scenes. Both FPVScore and the perceptual study benchmark prompt adherence (PA), layout correctness (LC), visual quality (VQ), style consistency (SC), and overall preference (OP). The best results are highlighted in bold . * denotes the retrieval mode.

Figure 4. Qualitative comparison. The first column shows the input multimodal graph with key scene edges. Red rectangles show inconsistencies and artifacts, while green rectangles highlight consistent regions. † Text-only scene graphs are given.

Figure 5. Qualitative comparison with training-free, language-based methods. The grey boxes display the input textual instructions derived from text-only scene graphs, while FlowScene takes the text-only graphs as input.

Method	Layout	Shape	Texture	Time (s)
EchoScene ( Zhai et al., 2024b )	✓	✓	–	44.27
MMGDreamer ( Yang et al., 2025a )	✓	✓	–	45.34
FlowScene (Ours, w/o TB)	✓	✓	✗	6.83
FlowScene (Ours)	✓	✓	✓	37.38

Table 4. Inference Time Comparison.

LEU	SEU	TEU	FID	$\text{FID}_{\text{CLIP}}$	KID
			50.83	4.55	7.32
✓			40.55	3.48	2.58
	✓		44.73	4.42	4.97
		✓	43.82	3.86	3.68
✓	✓		36.81	2.92	1.31
✓		✓	36.30	2.90	0.77

Table 5. Ablation Study on the InfoExchangeUnit across different branches, including the LayoutExchangeUnit (LEU), ShapeExchangeUnit (SEU), and TextureExchangeUnit (TEU).

Shape		Layout		FID	$\text{FID}_{\text{CLIP}}$	KID
Diffusion	Flow	Diffusion	Flow	FID	$\text{FID}_{\text{CLIP}}$	KID
✓		✓		35.23	3.26	1.78
✓			✓	34.96	2.90	1.03
	✓	✓		33.03	2.94	0.62
	✓		✓	32.76	2.86	0.38

Table 6. Ablation studies on different generation backbones.

Modality		FPVScore $\uparrow$
Visual	Textual	LC	VQ	SC	OP
10%	90%	3.28	3.37	3.42	3.26
40%	60%	3.45	3.86	3.89	3.67
50%	50%	3.42	3.88	3.91	3.64
60%	40%	3.36	3.88	3.91	3.62
90%	10%	3.28	3.78	3.82	3.59

Table 7. Ablation studies on the ratio of node modalities.

Appendix figures & tables13 assets

Supplementary material from the paper’s appendix.

Appendix

Method	Mode	left/ right	front/ behind	smaller/ larger	taller/ shorter	close by	symmetrical
Graph-to-3D ( Dhamo et al., 2021 )	Change	0.91	0.92	0.86	0.89	0.69	0.46
CommonScenes ( Zhai et al., 2024c )		0.91	0.92	0.86	0.91	0.69	0.59
EchoScene ( Zhai et al., 2024b )		0.94	0.96	0.92	0.93	0.74	0.50
MMGDreamer ( Yang et al., 2025a )		0.95	0.96	0.93	0.93	0.71	0.53
FlowScene (Ours)		0.95	0.96	0.93	0.94	0.75	0.58
Graph-to-3D ( Dhamo et al., 2021 )	Addition	0.94	0.95	0.91	0.93	0.63	0.47

Appendix

Table 8. Scene graph constraints. Comparisons are across three modes: Relationship Change, Node Addition, and Generation Only (listed top to bottom). The best results are highlighted in bold .

Method	Bedroom		Living room		Dining room
Method	N.stand	Lamp	Table	Lamp	Chair	Table
CommonScenes ( Zhai et al., 2024c )	2.69	-	11.75	-	1.96	9.04
EchoScene ( Zhai et al., 2024b )	1.68	30.07	3.02	10.06	1.75	1.26
MMGDreamer ( Yang et al., 2025a )	1.33	2.29	2.44	0.18	0.23	0.83
FlowScene (Ours)	0.46	0.19	1.61	0.14	0.21	0.37

Appendix

Table 9. Inter-shape consistency. The Chamfer Distance (CD,

\times 0.001

) is computed between object pairs linked by the same as relationship. Lower values indicate a higher degree of shape similarity between paired instances.

Training		Testing		FID	$\text{FID}_{\text{CLIP}}$	KID
Multi	Single	Multi	Single	FID	$\text{FID}_{\text{CLIP}}$	KID
✓		✓		33.77	2.88	0.73
✓			✓	33.79	2.97	0.99
	✓	✓		40.73	4.04	3.79
	✓		✓	32.76	2.86	0.38

Appendix

Table 10. Ablation studies of different input views.

Mask Ratio	FID	$\text{FID}_{\text{CLIP}}$	KID
20%	34.40	2.83	0.43
50%	38.94	3.22	0.96
80%	43.54	3.78	2.19
0% ∗	40.89	3.36	2.05

Appendix

Table 11. Ablation Study on the mask ratio of relation. * indicates results from MMGDreamer.

System & Hardware Overview
CPU	AMD EPYC 7702 64-Core
GPU	8 $\times$ NVIDIA A100 GPU
Memory	1T RAM
Operating System	Ubuntu 22.04 LTS
CUDA Version	12.8
NVIDIA Driver	570.133.20

Appendix

Table 12. System and Software Configuration.

Figure 7. Qualitative comparison with graph-conditioned generative models. The top row illustrates the input graphs, highlighting key semantic and spatial edges. In the generated scenes, red rectangles indicate generation failures in baselines (e.g., collisions), whereas green rectangles mark consistent regions in our results. * denotes the retrieval mode. † Text-only scene graphs are given.

Figure 8. Qualitative comparison on object generation with other methods.

Figure 9. Failure case. The left panel shows the input multimodal scene graph, while the right panel shows the generated failure case. Red cross marks indicate removed relationships.

Panel of Figure 10 — Figure 10. Prompt template for evaluating FPVScore ( Huang et al., 2025 ) with GPT-4o ( Achiam et al., 2023 ) .

Figure 11. Perceptual Study Interface: Instructions and Metrics. This introductory page is presented to participants at the start of the perceptual study. It outlines the study objectives, provides detailed instructions for viewing and rating, and clearly defines the five evaluation metrics (PA, LC, VQ, SC, OP).

Figure 12. Perceptual Study Interface: Example Page. Participants are presented with a detailed text prompt (top) and the corresponding generated 3D scenes from five different methods. Users interactively rate each method using sliding scales (1.00-10.00) across the five defined criteria, allowing for a direct and comprehensive comparison of generation quality.

Figure 13. The Interface of Language Description with Images.

Image-to-3D Scene Generation

See all 24 papers

DistScene: Object-to-Scene Distillation for 3D Scene Generation
Oct 3, 2026Kunming Luo, Hongyu Yan, Ken Deng +63D Scene Generation Image-to-3D Scene Generation
The Hong Kong University of Science and Technology · TeleAI, China Telecom
ScenePilot: Grow-and-Repair Policy for Text-Driven 3D Indoor Scene Generation
Aug 31, 2026Jiawei Zhang, Hongsong Wang, Pan Zhou3D Scene Editing Functional 3D Scene Generation
Southeast University · Singapore Management University
WorldMesh: Generating Navigable Multi-Room 3D Scenes via Mesh-Conditioned Image Diffusion
Mar 24, 2026Manuel-Andreas Schneider, Angela Dai3D Scene Generation 3D Scene Representation
Technical University of Munich, Germany

Oct 3, 2026J/K move · Enter open · S save

DistScene: Object-to-Scene Distillation for 3D Scene Generation

Kunming Luo, Hongyu Yan, Ken Deng, Chengcheng Zhou+5

The Hong Kong University of Science and Technology · TeleAI, China Telecom

3D Scene Generation Image-to-3D Scene Generation 3D Scene Layout Generation Functional 3D Scene Generation Knowledge Distillation

Open paper AskPDF arXiv

We present DistScene, a framework for single-image compositional 3D scene generation by jointly modeling the environment and individual objects. Unlike existing methods that represent scenes primarily as collections of objects, we model the environment as an explicit scene component to provide geometric context for object placement. Specifically, we introduce Scene-Frame Generation, which jointly generates separate environment and object components in a shared coordinate frame, allowing their geometry and relative placement to be learned together. Then we introduce Object-Centric Refinement to refine each object in a local frame with scene context. Finally, we develop Object-to-Scene Distillation to transfer pretrained object-generation priors to scene generation through automatically composed and rendered synthetic scenes. Evaluations on indoor and outdoor benchmarks demonstrate improved scene-level spatial coherence over the evaluated baselines. Project page: https://coolbeam.github.io/DistScene/

Abstract

Image-to-3D Scene Generation

Explore similar work

DistScene: Object-to-Scene Distillation for 3D Scene Generation

ScenePilot: Grow-and-Repair Policy for Text-Driven 3D Indoor Scene Generation

WorldMesh: Generating Navigable Multi-Room 3D Scenes via Mesh-Conditioned Image Diffusion