SpatialEdit-Bench

SpatialEdit-Bench is a comprehensive benchmark dataset for fine-grained image spatial editing, introduced in the paper SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing.

It is designed to evaluate whether an editing model can produce visually plausible results while accurately following spatial transformation instructions. The benchmark covers both object-centric and camera-centric editing tasks, focusing on precise spatial control rather than just appearance changes.

Resources

• GitHub Repository: EasonXiao-888/SpatialEdit

Evaluation Scope

SpatialEdit-Bench is built to assess edits such as:

• Object Moving: Relocating objects within a scene.
• Object Rotation: Precise changes in object orientation.
• Camera Viewpoint Change: Systematic transformations of the camera trajectory and framing.

Usage

Please refer to the GitHub repository for environment setup and further evaluation scripts (camera-level and object-level).

Citation

@article{xiao2026spatialedit,
  title   = {SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing},
  author  = {Xiao, Yicheng and Zhang, Wenhu and Song, Lin and Chen, Yukang and Li, Wenbo and Jiang, Nan and Ren, Tianhe and Lin, Haokun and Huang, Wei and Huang, Haoyang and Li, Xiu and Duan, Nan and Qi, Xiaojuan},
  journal = {arXiv preprint arXiv:2604.04911},
  year    = {2026}
}