PulseAugur
实时 20:34:32
English(EN) SEAL: Semantic-aware Single-image Sticker Personalization with a Large-scale Sticker-tag Dataset

SEAL通过解决过拟合和结构刚性问题,改进了AI贴纸个性化

研究人员开发了SEAL,一种使用单个参考图像对文本到图像生成中的贴纸进行个性化的新方法。SEAL解决了现有测试时微调方法中出现的视觉纠缠和结构刚性等问题。该方法可作为即插即用模块集成到扩散模型中,并利用了语义引导空间注意力损失、拆分合并令牌策略和结构感知层限制。为了便于评估,创建了一个名为StickerBench的大规模数据集,其中包含用于属性级控制的结构化标签。 AI

影响 引入了一种新颖的个性化图像生成技术,有望提高贴纸创建的控制力并减少伪影。

排序理由 这是一篇描述新方法和图像生成数据集的研究论文。

在 arXiv cs.CV 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

SEAL通过解决过拟合和结构刚性问题,改进了AI贴纸个性化

报道来源 [2]

  1. arXiv cs.CV TIER_1 English(EN) · Changhyun Roh, Yonghyun Jeong, Jonghyun Lee, Chanho Eom, Jihyong Oh ·

    SEAL: Semantic-aware Single-image Sticker Personalization with a Large-scale Sticker-tag Dataset

    arXiv:2604.26883v1 Announce Type: new Abstract: Synthesizing a target concept from a single reference image is challenging in diffusion-based personalized text-to-image generation, particularly for sticker personalization where prompts often require explicit attribute edits. With…

  2. arXiv cs.CV TIER_1 English(EN) · Jihyong Oh ·

    SEAL: Semantic-aware Single-image Sticker Personalization with a Large-scale Sticker-tag Dataset

    Synthesizing a target concept from a single reference image is challenging in diffusion-based personalized text-to-image generation, particularly for sticker personalization where prompts often require explicit attribute edits. With only one reference, test-time fine-tuning (TTF)…