# InsertFuse: A Unified Framework for Multi-Category Reference-Guided Image Insertion

arXiv: https://arxiv.org/abs/2608.06490 · HTML: https://mlsanity.com/paper/2608.06490/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 37 references that are in the map.
- First posted on arXiv 6 Aug 2026.

## Builds on (most cited first, 25 of 25)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Prompt-to-Prompt Image Editing with Cross Attention Control](https://mlsanity.com/paper/2208.01626/) (2022)
- [DiffEdit: Diffusion-based semantic image editing with mask guidance](https://mlsanity.com/paper/2210.11427/) (2022)
- [MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editing](https://mlsanity.com/paper/2306.10012/) (2023)
- [Paint by Example: Exemplar-based Image Editing with Diffusion Models](https://mlsanity.com/paper/2211.13227/) (2022)
- [InstantID: Zero-shot Identity-Preserving Generation in Seconds](https://mlsanity.com/paper/2401.07519/) (2024)
- [Step1X-Edit: A Practical Framework for General Image Editing](https://mlsanity.com/paper/2504.17761/) (2025)
- [PhotoMaker: Customizing Realistic Human Photos via Stacked ID Embedding](https://mlsanity.com/paper/2312.04461/) (2023)
- [Emu Edit: Precise Image Editing via Recognition and Generation Tasks](https://mlsanity.com/paper/2311.10089/) (2023)
- [Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding](https://mlsanity.com/paper/2405.08748/) (2024)
- [AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea](https://mlsanity.com/paper/2411.15738/) (2024)
- [In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer](https://mlsanity.com/paper/2504.20690/) (2025)
- [StableVITON: Learning Semantic Correspondence with Latent Diffusion Model for Virtual Try-On](https://mlsanity.com/paper/2312.01725/) (2023)
- [Improving Diffusion Models for Authentic Virtual Try-on in the Wild](https://mlsanity.com/paper/2403.05139/) (2024)
- [CatVTON: Concatenation Is All You Need for Virtual Try-On with Diffusion Models](https://mlsanity.com/paper/2407.15886/) (2024)
- [Lumina-Image 2.0: A Unified and Efficient Image Generative Framework](https://mlsanity.com/paper/2503.21758/) (2025)
- [ACE++: Instruction-Based Image Creation and Editing via Context-Aware Content Filling](https://mlsanity.com/paper/2501.02487/) (2025)
- [CogView3: Finer and Faster Text-to-Image Generation via Relay Diffusion](https://mlsanity.com/paper/2403.05121/) (2024)
- [Insert Anything: Image Insertion via In-Context Editing in DiT](https://mlsanity.com/paper/2504.15009/) (2025)
- [Qwen-Image-2.0 Technical Report](https://mlsanity.com/paper/2605.10730/) (2026)
- [ObjectMate: A Recurrence Prior for Object Insertion and Subject-Driven Generation](https://mlsanity.com/paper/2412.08645/) (2024)
- [BIFR"OST: 3D-Aware Image compositing with Language Instructions](https://mlsanity.com/paper/2410.19079/) (2024)
- [Direct 3D-Aware Object Insertion via Decomposed Visual Proxies](https://mlsanity.com/paper/2606.06601/) (2026)
- [A2-Edit: Precise Reference-Guided Image Editing of Arbitrary Objects and Ambiguous Masks](https://mlsanity.com/paper/2603.10685/) (2026)

## Built on by (most cited first, 1 of 1)

- [Depth-to-RGB: Repurposing a Frozen Depth Estimator for Geometry-Guided Compositing](https://mlsanity.com/paper/2610.09125/) (2026)

## Abstract (arXiv)

We present InsertFuse, a unified framework for multi-category reference-guided image insertion. Its key idea is to decouple category-specific expertise learning from cross-category capability consolidation. InsertFuse first trains specialized experts for different insertion categories and then introduces Insertion On-Policy Distillation (IOPD) to consolidate their capabilities into a single student. By querying the matched expert at states visited by the student, IOPD preserves category-specific insertion behavior while mitigating the cross-category interference caused by direct joint training. To improve spatial control, we propose Token-Aligned Geometry Conditioning (TAGC), which maps mask-derived geometric cues to the visual token grid, and Region-Balanced Flow Matching, which separately normalizes prediction errors inside and outside the insertion region to prevent background-dominated and scale-dependent supervision. We further introduce Reference CFG to isolate and strengthen the guidance induced by the visual reference under fixed scene and geometry conditions, with IOPD transferring this enhanced supervision into the unified student. Extensive experiments on the public AnyInsertion benchmark and our multi-category test set demonstrate state-of-the-art performance on most metrics, showing strong reference fidelity and generation quality across diverse insertion categories.