Saved in:
Bibliographic Details
Main Authors: Guo, Hanzhong, Yu, Yizhou
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.20807
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918513613996032
author Guo, Hanzhong
Yu, Yizhou
author_facet Guo, Hanzhong
Yu, Yizhou
contents Subject-driven text-to-image generation still struggles to preserve high-frequency identity details such as logos, patterns, and text. Existing methods typically operate directly in RGB space, which often leads to detail degradation under substantial edits. We propose a two-stage framework that decouples structure from appearance by first predicting a Canny map and then rendering the final image conditioned on both the source appearance and the predicted structure. To improve text handling, we further introduce a fully automatic pipeline that constructs a 100k-pair text-aware dataset with cross-view textual consistency. Experiments, including GPT-4.1-based evaluation and a knowledge distillation study, show clear gains over selected baselines and suggest that intermediate structural prediction is an effective route for high-fidelity subject-driven generation. Our dataset and code will be made publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2605_20807
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Decomposing Subject-Driven Image Generation via Intermediate Structural Prediction
Guo, Hanzhong
Yu, Yizhou
Computer Vision and Pattern Recognition
Subject-driven text-to-image generation still struggles to preserve high-frequency identity details such as logos, patterns, and text. Existing methods typically operate directly in RGB space, which often leads to detail degradation under substantial edits. We propose a two-stage framework that decouples structure from appearance by first predicting a Canny map and then rendering the final image conditioned on both the source appearance and the predicted structure. To improve text handling, we further introduce a fully automatic pipeline that constructs a 100k-pair text-aware dataset with cross-view textual consistency. Experiments, including GPT-4.1-based evaluation and a knowledge distillation study, show clear gains over selected baselines and suggest that intermediate structural prediction is an effective route for high-fidelity subject-driven generation. Our dataset and code will be made publicly available.
title Decomposing Subject-Driven Image Generation via Intermediate Structural Prediction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.20807