PosterReward: Unlocking Accurate Evaluation for High-Quality Graphic Design Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914435415670784 |
|---|---|
| author | Lai, Jianyu Chen, Sixiang Gao, Jialin Shi, Hengyu Liu, Zhongying Zhai, Fuxiang Luo, Junfeng Wei, Xiaoming Wang, Lujia Zhu, Lei |
| author_facet | Lai, Jianyu Chen, Sixiang Gao, Jialin Shi, Hengyu Liu, Zhongying Zhai, Fuxiang Luo, Junfeng Wei, Xiaoming Wang, Lujia Zhu, Lei |
| contents | Recent advancements in the text-rendering capabilities of image generation models have made the end-to-end creation of graphic design content, such as posters, increasingly feasible. However, existing reward models fall short of accurately assessing design quality, as they primarily focus on global image aesthetics while overlooking the critical dimensions of typography and layout. Furthermore, the scarcity of domain-specific preference data remains a significant bottleneck, which limits the further development of graphic design evaluation and generation.
To bridge this gap, we introduce an automated pipeline to construct a high-quality dataset of 70k poster preferences by leveraging the consensus of multiple Multi-modal Large Language Models (MLLMs) to simulate human-like judgment. Utilizing this dataset, we develop PosterReward, a reward model specifically designed for high-precision poster assessment through a cascaded, multi-stage training strategy. We also provide multiple variants of the model to cater to different application scenarios. Finally, we introduce PosterRewardBench and PosterBench to evaluate the performance of existing reward models in poster assessment and the generation capabilities of current text-to-image models in poster creation, respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_29855 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PosterReward: Unlocking Accurate Evaluation for High-Quality Graphic Design Generation Lai, Jianyu Chen, Sixiang Gao, Jialin Shi, Hengyu Liu, Zhongying Zhai, Fuxiang Luo, Junfeng Wei, Xiaoming Wang, Lujia Zhu, Lei Graphics Recent advancements in the text-rendering capabilities of image generation models have made the end-to-end creation of graphic design content, such as posters, increasingly feasible. However, existing reward models fall short of accurately assessing design quality, as they primarily focus on global image aesthetics while overlooking the critical dimensions of typography and layout. Furthermore, the scarcity of domain-specific preference data remains a significant bottleneck, which limits the further development of graphic design evaluation and generation. To bridge this gap, we introduce an automated pipeline to construct a high-quality dataset of 70k poster preferences by leveraging the consensus of multiple Multi-modal Large Language Models (MLLMs) to simulate human-like judgment. Utilizing this dataset, we develop PosterReward, a reward model specifically designed for high-precision poster assessment through a cascaded, multi-stage training strategy. We also provide multiple variants of the model to cater to different application scenarios. Finally, we introduce PosterRewardBench and PosterBench to evaluate the performance of existing reward models in poster assessment and the generation capabilities of current text-to-image models in poster creation, respectively. |
| title | PosterReward: Unlocking Accurate Evaluation for High-Quality Graphic Design Generation |
| topic | Graphics |
| url | https://arxiv.org/abs/2603.29855 |