# BoxCtrl: 3D-Aware Visual Prompting for Geometric Image Editing

arXiv: https://arxiv.org/abs/2606.23270 · HTML: https://mlsanity.com/paper/2606.23270/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Embodied and 3D vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 28 of its 38 references that are in the map.
- First posted on arXiv 22 Jun 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 28 of 28)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [Qwen-Image Technical Report](https://mlsanity.com/paper/2508.02324/) (2025)
- [Grounded SAM: Assembling Open-World Models for Diverse Visual Tasks](https://mlsanity.com/paper/2401.14159/) (2024)
- [Google Scanned Objects: A High-Quality Dataset of 3D Scanned Household Items](https://mlsanity.com/paper/2204.11918/) (2022)
- [AnyDoor: Zero-shot Object-level Image Customization](https://mlsanity.com/paper/2307.09481/) (2023)
- [OminiControl: Minimal and Universal Control for Diffusion Transformer](https://mlsanity.com/paper/2411.15098/) (2024)
- [LooseControl: Lifting ControlNet for Generalized Depth Conditioning](https://mlsanity.com/paper/2312.03079/) (2023)
- [OBJECT 3DIT: Language-guided 3D-aware Image Editing](https://mlsanity.com/paper/2307.11073/) (2023)
- [Diffusion Handles: Enabling 3D Edits for Diffusion Models by Lifting Activations to 3D](https://mlsanity.com/paper/2312.02190/) (2023)
- [Image Sculpting: Precise Object Editing with 3D Geometry Control](https://mlsanity.com/paper/2401.01702/) (2024)
- [Magic Fixup: Streamlining Photo Editing by Watching Dynamic Videos](https://mlsanity.com/paper/2403.13044/) (2024)
- [GeoDiffuser: Geometry-Based Image Editing with Diffusion Models](https://mlsanity.com/paper/2404.14403/) (2024)
- [Detect Anything 3D in the Wild](https://mlsanity.com/paper/2504.07958/) (2025)
- [Neural Assets: 3D-Aware Multi-Object Scene Synthesis with Image Diffusion Models](https://mlsanity.com/paper/2406.09292/) (2024)
- [ObjectMover: Generative Object Movement with Video Prior](https://mlsanity.com/paper/2503.08037/) (2025)
- [Orient Anything V2: Unifying Orientation and Rotation Understanding](https://mlsanity.com/paper/2601.05573/) (2026)
- [Build-A-Scene: Interactive 3D Layout Control for Diffusion-Based Image Generation](https://mlsanity.com/paper/2408.14819/) (2024)
- [VideoHandles: Editing 3D Object Compositions in Videos Using Video Generative Priors](https://mlsanity.com/paper/2503.01107/) (2025)
- [Customizing Text-to-Image Diffusion with Object Viewpoint Control](https://mlsanity.com/paper/2404.12333/) (2024)
- [Training-free Geometric Image Editing on Diffusion Models](https://mlsanity.com/paper/2507.23300/) (2025)
- [3D-Fixup: Advancing Photo Editing with 3D Priors](https://mlsanity.com/paper/2505.10566/) (2025)
- [I2V3D: Controllable image-to-video generation with 3D guidance](https://mlsanity.com/paper/2503.09733/) (2025)
- [BlenderFusion: 3D-Grounded Visual Editing and Generative Compositing](https://mlsanity.com/paper/2506.17450/) (2025)
- [BlobCtrl: Taming Controllable Blob for Element-level Image Editing](https://mlsanity.com/paper/2503.13434/) (2025)
- [Controllable 3D Placement of Objects with Scene-Aware Diffusion Models](https://mlsanity.com/paper/2506.21446/) (2025)

## Abstract (arXiv)

As instruction-based editing models and multimodal large language models advance, diverse image editing tasks have become feasible. However, achieving precise and consistent geometric image editing, such as translating, scaling, and rotating in 3D space, remains a major challenge. In this work, we introduce BoxCtrl, a 3D-aware visual prompting framework. Unlike text-only or coarse 2D-guided approaches, our method introduces informative RGB 3D bounding boxes projected onto 2D images as visual prompts. The three orthogonal faces of each box are painted with distinct RGB colors, simultaneously encoding position, size, and orientation to provide a compact, intuitive in-context visual example. The key to BoxCtrl's success lies in these well-designed bounding boxes, which decouple geometric control from appearance control. This enables the model to learn consistent correspondences between faces of the same color in the latent space, leading to a precise understanding of geometric intentions and accurate editing results. We introduce a two-stage training paradigm: Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL). To address paired data scarcity, we construct a large-scale synthetic dataset for SFT, equipping the model with fundamental editing capabilities. To bridge the synthetic-to-real domain gap, we incorporate an online RL stage leveraging unpaired real-world data. Guided by a reward function evaluating geometric accuracy and visual fidelity, our SFT-RL strategy significantly enhances geometric precision while maintaining photorealistic quality. Extensive experiments demonstrate that BoxCtrl achieves state-of-the-art performance across translation, rotation, scaling, and composite editing tasks.