Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.14638 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908545406992384 |
|---|---|
| author | Li, Mingsong Liu, Lin Wang, Hongjun Chen, Haoxing Gu, Xijun Liu, Shizhan Gong, Dong Zhao, Junbo Lan, Zhenzhong Li, Jianguo |
| author_facet | Li, Mingsong Liu, Lin Wang, Hongjun Chen, Haoxing Gu, Xijun Liu, Shizhan Gong, Dong Zhao, Junbo Lan, Zhenzhong Li, Jianguo |
| contents | Current instruction-based image editing (IBIE) methods struggle with challenging editing tasks, as both editing types and sample counts of existing datasets are limited. Moreover, traditional dataset construction often contains noisy image-caption pairs, which may introduce biases and limit model capabilities in complex editing scenarios. To address these limitations, we introduce MultiEdit, a comprehensive dataset featuring over 107K high-quality image editing samples. It encompasses 6 challenging editing tasks through a diverse collection of 18 non-style-transfer editing types and 38 style transfer operations, covering a spectrum from sophisticated style transfer to complex semantic operations like person reference editing and in-image text editing. We employ a novel dataset construction pipeline that utilizes two multi-modal large language models (MLLMs) to generate visual-adaptive editing instructions and produce high-fidelity edited images, respectively. Extensive experiments demonstrate that fine-tuning foundational open-source models with our MultiEdit-Train set substantially improves models' performance on sophisticated editing tasks in our proposed MultiEdit-Test benchmark, while effectively preserving their capabilities on the standard editing benchmark. We believe MultiEdit provides a valuable resource for advancing research into more diverse and challenging IBIE capabilities. Our dataset is available at https://huggingface.co/datasets/inclusionAI/MultiEdit. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_14638 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MultiEdit: Advancing Instruction-based Image Editing on Diverse and Challenging Tasks Li, Mingsong Liu, Lin Wang, Hongjun Chen, Haoxing Gu, Xijun Liu, Shizhan Gong, Dong Zhao, Junbo Lan, Zhenzhong Li, Jianguo Computer Vision and Pattern Recognition Current instruction-based image editing (IBIE) methods struggle with challenging editing tasks, as both editing types and sample counts of existing datasets are limited. Moreover, traditional dataset construction often contains noisy image-caption pairs, which may introduce biases and limit model capabilities in complex editing scenarios. To address these limitations, we introduce MultiEdit, a comprehensive dataset featuring over 107K high-quality image editing samples. It encompasses 6 challenging editing tasks through a diverse collection of 18 non-style-transfer editing types and 38 style transfer operations, covering a spectrum from sophisticated style transfer to complex semantic operations like person reference editing and in-image text editing. We employ a novel dataset construction pipeline that utilizes two multi-modal large language models (MLLMs) to generate visual-adaptive editing instructions and produce high-fidelity edited images, respectively. Extensive experiments demonstrate that fine-tuning foundational open-source models with our MultiEdit-Train set substantially improves models' performance on sophisticated editing tasks in our proposed MultiEdit-Test benchmark, while effectively preserving their capabilities on the standard editing benchmark. We believe MultiEdit provides a valuable resource for advancing research into more diverse and challenging IBIE capabilities. Our dataset is available at https://huggingface.co/datasets/inclusionAI/MultiEdit. |
| title | MultiEdit: Advancing Instruction-based Image Editing on Diverse and Challenging Tasks |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.14638 |