Draw-In-Mind: Rebalancing Designer-Painter Roles in Unified Multimodal Models Benefits Image Editing
This work addresses the challenge of precise image editing for users of multimodal AI models, offering an incremental improvement by rebalancing module roles rather than introducing a new paradigm.
The paper tackles the problem of precise image editing in unified multimodal models by addressing an imbalanced division of responsibilities between understanding and generation modules. It introduces Draw-In-Mind (DIM), a dataset with 14M image-text pairs and 233K chain-of-thought imaginations, and trains a model that achieves state-of-the-art or competitive performance on benchmarks like ImgEdit and GEdit-Bench, outperforming larger models such as UniWorld-V1 and Step1X-Edit.
In recent years, integrating multimodal understanding and generation into a single unified model has emerged as a promising paradigm. While this approach achieves strong results in text-to-image (T2I) generation, it still struggles with precise image editing. We attribute this limitation to an imbalanced division of responsibilities. The understanding module primarily functions as a translator that encodes user instructions into semantic conditions, while the generation module must simultaneously act as designer and painter, inferring the original layout, identifying the target editing region, and rendering the new content. This imbalance is counterintuitive because the understanding module is typically trained with several times more data on complex reasoning tasks than the generation module. To address this issue, we introduce Draw-In-Mind (DIM), a dataset comprising two complementary subsets: (i) DIM-T2I, containing 14M long-context image-text pairs to enhance complex instruction comprehension; and (ii) DIM-Edit, consisting of 233K chain-of-thought imaginations generated by GPT-4o, serving as explicit design blueprints for image edits. We connect a frozen Qwen2.5-VL-3B with a trainable SANA1.5-1.6B via a lightweight two-layer MLP, and train it on the proposed DIM dataset, resulting in DIM-4.6B-T2I/Edit. Despite its modest parameter scale, DIM-4.6B-Edit achieves SOTA or competitive performance on the ImgEdit and GEdit-Bench benchmarks, outperforming much larger models such as UniWorld-V1 and Step1X-Edit. These findings demonstrate that explicitly assigning the design responsibility to the understanding module provides significant benefits for image editing. Our dataset and models are available at https://github.com/showlab/DIM.