Lune

CVPR2026顶会

Foundation Encoders Are All You Need for Preference-Aware Personalization

Hyungjin Kim, Seokho Ahn, Young-Duk Seo

出版方
2026年份

摘要

Bicycle on bridge Swirling a wine glass Dog running through fog River bending through sky Empty bench Minimal line drawing, fine ink lines, … User 2 Silhouette of a loving couple

Figure 1. FAN personalizes synthesis results on foundation encoders without additional structure or fine-tuning, integrating key interests (e.g., style, texture, color) within the latent space to satisfy individual preference. Personalized results using a single style prompt on FLUX.

问问这篇 Paper

智能体会读完全文。

Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。

可以从这些问题问起

智能体调用

Luneget_paper_fulltext

在 Lune 里问

免费开始,无需绑卡

它引用的顶会 Paper20

相关 Paper

黄昏的海面,两侧是细线勾勒的悬崖