cs.CVNov 1, 2026

Restoring Semantic Fidelity for Training-Free High-Resolution Image Generation in Diffusion Transformers

Authors: Junjie Li, Zengbin Wang, Gaoxiang Li, Muyi Sun, Yixin Zhang, Man Zhang

Organizations: Beijing University of Posts and Telecommunications, Haidian, Beijing, China · Beijing Normal University, Haidian, Beijing, China · Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, Hefei, Anhui, China · Qinghai Institute of Technology, Xining, Qinghai, China

Abstract

Diffusion Transformer (DiT)-based text-to-image models have recently achieved remarkable success in image synthesis. However, when directly applied beyond their training resolution, these models often suffer from severe semantic degradation, where meaningful semantic content disappears. Although existing training-free high-resolution generation methods can mitigate this issue, the inherent semantic degradation at high resolutions remains largely unresolved, leading to degraded fine-grained details. Furthermore, widely used structural guidance methods lack modeling of the structure and details in the intermediate trajectory, which degrades the quality of the generated results. In this paper, we propose a training-free framework for high-resolution image generation to address these challenges. First, we propose a text-visual scale-decoupled attention (TVSDA) module, which restores the model's ability to generate semantically meaningful content at high resolutions by preserving the global receptive field of text tokens while aligning the receptive field of image tokens with the training resolution. Second, we propose a structure-detail guidance (SDG) strategy that explicitly models structural information and fine details through scale-wise and temporal-wise guidance, overcoming the limitations of conventional low-resolution guidance. Extensive experiments demonstrate that our method significantly improves both structural coherence and detail fidelity over existing training-free DiT-based high-resolution generation approaches. The code is available at https://github.com/ljjcoder/Restoring-Semantic-Fidelity.

Figures & tables

Explore similar work

CardsList
  1. HRDiT: Training-Free High-Resolution Image Generation with Off-the-Shelf Diffusion Transformer Models

    Aug 7, 2026Yu Xue, Haoxuan Qu, Zhuoling Li +5Diffusion Model AdaptationDiffusion Transformer

  2. SEGA: Spectral-Energy Guided Attention for Resolution Extrapolation in Diffusion Transformers

    May 21, 2026Javad Rajabi, Kimia Shaban, Koorosh Roohi +2Rotary Positional EmbeddingsDiffusion Model Guidance

  3. TIDE: Text-Informed Dynamic Extrapolation with Step-Aware Temperature Control for Diffusion Transformers

    Mar 9, 2026Yihua Liu, Fanjiang Ye, Bowen Lin +2T2I Diffusion ModelsDiffusion Transformer