SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing

Yicheng Xiao; Wenhu Zhang; Lin Song; Yukang Chen; Wenbo Li; Nan Jiang; Tianhe Ren; Haokun Lin

cs.CVApr 6, 2026

SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing

Authors: Yicheng Xiao, Wenhu Zhang, Lin Song, Yukang Chen, Wenbo Li, Nan Jiang, Tianhe Ren, Haokun Lin,

Organizations: The University of Hong Kong · JD Explore Academy · Tsinghua University · The Hong Kong University of Science and Technology ·

Abstract

Image spatial editing performs geometry-driven transformations, allowing precise control over object layout and camera viewpoints. Current models are insufficient for fine-grained spatial manipulations, motivating a dedicated assessment suite. Our contributions are listed: (i) We introduce SpatialEdit-Bench, a complete benchmark that evaluates spatial editing by jointly measuring perceptual plausibility and geometric fidelity via viewpoint reconstruction and framing analysis. (ii) To address the data bottleneck for scalable training, we construct SpatialEdit-500k, a synthetic dataset generated with a controllable Blender pipeline that renders objects across diverse backgrounds and systematic camera trajectories, providing precise ground-truth transformations for both object- and camera-centric operations. (iii) Building on this data, we develop SpatialEdit-16B, a baseline model for fine-grained spatial editing. Our method achieves competitive performance on general editing while substantially outperforming prior methods on spatial manipulation tasks. All resources will be made public at https://github.com/EasonXiao-888/SpatialEdit.

Figures & tables

Figure 1: Illustration for image spatial editing . It comprises two components: (1) camera-centric view manipulation, including pitch, yaw, and zoom transformations; and (2) single-object manipulation, encompassing object rotation while preserving the background, as well as translation and scaling of objects specified via user-defined bounding boxes.

Benchmark	Object	Object	Object	Camera	VLM	Precise
Benchmark	Translation	Scaling	Rotation	Manipulation	Metric	Metric
ImgEdit [ 65 ]	✗	✗	✗	✗	✔	✗
GEdit [ 32 ]	✔	✗	✗	✗	✔	✗
CEdit [ 46 ]	✔	✗	✗	✔	✔	✗
SpatialEdit-Bench	✔	✔	✔	✔	✔	✔

Table 1: Characteristics comparison with other editing benchmarks.

Figure 2: Statistics of SpatialEdit-500k . (a) Distribution of camera-level data across seven sub-tasks in outdoor and intdoor scenes, where Y, P, and D denote Yaw, Pitch, and Distance, respectively. (b) Aspect ratio distribution of bounding boxes for the moving task at the object level. (c) Object category statistics across the entire dataset.

Figure 3: SpatialEdit-500k data generation pipeline . We leverage Blender to synthesize both objects and scenes, while preprocessing 3D assets using SAM3 and a vision-language model. The object-level engine constructs two inpainting-based data branches to generate object transformations, including rotation, translation, and scaling. The camera-level engine produces viewpoint transformation data by sampling different camera poses, resulting in variations in yaw, pitch, and zoom.

World Model
Method	Object		Camera		Object	Camera
	Moving	Rotation	Viewpoint	Framing	Overall	Overall
	Score $\uparrow$	Score $\uparrow$	Error $\downarrow$	Error $\downarrow$	Score $\uparrow$	Error $\downarrow$
ViduQ2-Turbo [ 50 ]	–	–	1.022	0.771	–	0.897
Kling-V2.5 [ 26 ]	–	–	1.051	0.733	–	0.892
Closed-Source Image Model

Table 2: Performance comparison on proposed SpatialEdit-Bench.

Figure 5: Comparison of camera view manipulation across various methods .

Figure 6: Comparison of object-level manipulation across various methods .

Table 9

Figure 7: Serving as an enhancement tool for single-view reconstruction .

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

World Model
Method	Camera		Camera
	Viewpoint	Framing	Overall
	Error $\downarrow$	Error $\downarrow$	Error $\downarrow$
Veo3.1 [ 17 ]	1.351	0.749	1.050
ViduQ2-Turbo [ 50 ]	1.022	0.771	0.897
Kling-V2.5 [ 26 ]	1.051	0.733	0.892

Appendix

Table 6: Performance comparison on proposed SpatialEdit-Bench.

Figure 8: Comparison of object-level manipulation across various methods .

Figure 9: Comparison of camera-level manipulation across various methods .

Image Editing

Image editing changes selected visual content while preserving the parts of an image that should remain unchanged.

See all 302 papers

3D-Layout-R1: Structured Reasoning for Language-Instructed Spatial Editing
Mar 23, 2026Haoyu Zhen, Xiaolong Li, Yilin Zhao +5Text-Guided 3D Scene Editing 3D Scene Editing
NVIDIA · UMass Amherst
Generative Visual Chain-of-Thought for Image Editing
Mar 2, 2026Zijin Yin, Tiankai Hang, Yiji Cheng +9Text-Guided Image Editing Visual CoT
Beijing University of Posts and Telecommunications · Beijing Key Laboratory of Multimodal Data Intelligent Perception and Governance · Tencent Hunyuan +1
InterCoG: Towards Spatially Precise Image Editing with Interleaved Chain-of-Grounding Reasoning
Mar 2, 2026Yecong Wan, Fan Li, Chunwei Wang +3Image Editing Text-Guided Image Editing
Faculty of Computing, Harbin Institute of Technology, Harbin, 150001, China. · Zhengzhou Advanced Research Institute of Harbin Institute of Technology, Zhengzhou, 450000, China. · Huawei Noah’s Ark Lab, Shenzhen, 518100, China. +1

Mar 23, 2026J/K move · Enter open · S save

3D-Layout-R1: Structured Reasoning for Language-Instructed Spatial Editing

Haoyu Zhen, Xiaolong Li, Yilin Zhao, Han Zhang+4

NVIDIA · UMass Amherst

Text-Guided 3D Scene Editing 3D Scene Editing 3D Spatial Reasoning LLM Reasoning with Graphs Spatial Reasoning Benchmarks

Open paper AskPDF arXiv

Large Language Models (LLMs) and Vision Language Models (VLMs) have shown impressive reasoning abilities, yet they struggle with spatial understanding and layout consistency when performing fine-grained visual editing. We introduce a Structured Reasoning framework that performs text-conditioned spatial layout editing via scene-graph reasoning. Given an input scene graph and a natural-language instruction, the model reasons over the graph to generate an updated scene graph that satisfies the text condition while maintaining spatial coherence. By explicitly guiding the reasoning process through structured relational representations, our approach improves both interpretability and control over spatial relationships. We evaluate our method on a new text-guided layout editing benchmark encompassing sorting, spatial alignment, and room-editing tasks. Our training paradigm yields an average 15% improvement in IoU and 25% reduction in center-distance error compared to Chain of Thought Fine-tuning (CoT-SFT) and vanilla GRPO baselines. Compared to SOTA zero-shot LLMs, our best models achieve up to 20% higher mIoU, demonstrating markedly improved spatial precision.

Abstract

Image Editing

Explore similar work

3D-Layout-R1: Structured Reasoning for Language-Instructed Spatial Editing

Generative Visual Chain-of-Thought for Image Editing

InterCoG: Towards Spatially Precise Image Editing with Interleaved Chain-of-Grounding Reasoning