Re-Align: Structured Reasoning-guided Alignment for In-Context Image Generation and Editing
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914241388216320 |
|---|---|
| author | He, Runze Cheng, Yiji Hang, Tiankai Li, Zhimin Xu, Yu Yin, Zijin Zhang, Shiyi Dai, Wenxun Du, Penghui Ma, Ao Wang, Chunyu Lu, Qinglin Han, Jizhong Dai, Jiao |
| author_facet | He, Runze Cheng, Yiji Hang, Tiankai Li, Zhimin Xu, Yu Yin, Zijin Zhang, Shiyi Dai, Wenxun Du, Penghui Ma, Ao Wang, Chunyu Lu, Qinglin Han, Jizhong Dai, Jiao |
| contents | In-context image generation and editing (ICGE) enables users to specify visual concepts through interleaved image-text prompts, demanding precise understanding and faithful execution of user intent. Although recent unified multimodal models exhibit promising understanding capabilities, these strengths often fail to transfer effectively to image generation. We introduce Re-Align, a unified framework that bridges the gap between understanding and generation through structured reasoning-guided alignment. At its core lies the In-Context Chain-of-Thought (IC-CoT), a structured reasoning paradigm that decouples semantic guidance and reference association, providing clear textual target and mitigating confusion among reference images. Furthermore, Re-Align introduces an effective RL training scheme that leverages a surrogate reward to measure the alignment between structured reasoning text and the generated image, thereby improving the model's overall performance on ICGE tasks. Extensive experiments verify that Re-Align outperforms competitive methods of comparable model scale and resources on both in-context image generation and editing tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_05124 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Re-Align: Structured Reasoning-guided Alignment for In-Context Image Generation and Editing He, Runze Cheng, Yiji Hang, Tiankai Li, Zhimin Xu, Yu Yin, Zijin Zhang, Shiyi Dai, Wenxun Du, Penghui Ma, Ao Wang, Chunyu Lu, Qinglin Han, Jizhong Dai, Jiao Computer Vision and Pattern Recognition In-context image generation and editing (ICGE) enables users to specify visual concepts through interleaved image-text prompts, demanding precise understanding and faithful execution of user intent. Although recent unified multimodal models exhibit promising understanding capabilities, these strengths often fail to transfer effectively to image generation. We introduce Re-Align, a unified framework that bridges the gap between understanding and generation through structured reasoning-guided alignment. At its core lies the In-Context Chain-of-Thought (IC-CoT), a structured reasoning paradigm that decouples semantic guidance and reference association, providing clear textual target and mitigating confusion among reference images. Furthermore, Re-Align introduces an effective RL training scheme that leverages a surrogate reward to measure the alignment between structured reasoning text and the generated image, thereby improving the model's overall performance on ICGE tasks. Extensive experiments verify that Re-Align outperforms competitive methods of comparable model scale and resources on both in-context image generation and editing tasks. |
| title | Re-Align: Structured Reasoning-guided Alignment for In-Context Image Generation and Editing |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.05124 |