Saved in:
Bibliographic Details
Main Authors: Lee, Kwanyoung, Oh, Hyunwoo, Cha, SeungJu, Koh, Sungho, Kim, Dong-Jin
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.19157
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911530537189376
author Lee, Kwanyoung
Oh, Hyunwoo
Cha, SeungJu
Koh, Sungho
Kim, Dong-Jin
author_facet Lee, Kwanyoung
Oh, Hyunwoo
Cha, SeungJu
Koh, Sungho
Kim, Dong-Jin
contents Generating rare compositional concepts in text-to-image synthesis remains a challenge for diffusion models, particularly for attributes that are uncommon in the training data. While recent approaches, such as R2F, address this challenge by utilizing LLM for prompt scheduling, they suffer from inherent variance due to the randomness of language models and suboptimal guidance from iterative text embedding switching. To address these problems, we propose the ADAPT framework, a training-free framework that deterministically plans and semantically aligns prompt schedules, providing consistent guidance to enhance the composition of rare concepts. By leveraging attention scores and orthogonal components, ADAPT significantly enhances compositional generation of rare concepts in the RareBench benchmark without additional training or fine-tuning. Through comprehensive experiments, we demonstrate that ADAPT achieves superior performance in RareBench and accurately reflects the semantic information of rare attributes, providing deterministic and precise control over the generation of rare compositions without compromising visual integrity.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19157
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ADAPT: Attention Driven Adaptive Prompt Scheduling and InTerpolating Orthogonal Complements for Rare Concepts Generation
Lee, Kwanyoung
Oh, Hyunwoo
Cha, SeungJu
Koh, Sungho
Kim, Dong-Jin
Computer Vision and Pattern Recognition
Generating rare compositional concepts in text-to-image synthesis remains a challenge for diffusion models, particularly for attributes that are uncommon in the training data. While recent approaches, such as R2F, address this challenge by utilizing LLM for prompt scheduling, they suffer from inherent variance due to the randomness of language models and suboptimal guidance from iterative text embedding switching. To address these problems, we propose the ADAPT framework, a training-free framework that deterministically plans and semantically aligns prompt schedules, providing consistent guidance to enhance the composition of rare concepts. By leveraging attention scores and orthogonal components, ADAPT significantly enhances compositional generation of rare concepts in the RareBench benchmark without additional training or fine-tuning. Through comprehensive experiments, we demonstrate that ADAPT achieves superior performance in RareBench and accurately reflects the semantic information of rare attributes, providing deterministic and precise control over the generation of rare compositions without compromising visual integrity.
title ADAPT: Attention Driven Adaptive Prompt Scheduling and InTerpolating Orthogonal Complements for Rare Concepts Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.19157