Jul 2026· International Journal of Computer Vision· Vol 134· 0 citations· 64 references
Computer Science
TL;DR
This training-free method improves the performance of both U-Net-based and Transformer-based diffusion models, including the Stable Diffusion series and FLUX series and adapts self-cross guidance as an effective reward for RL-based post-training and show improved subject diversity with no computational overhead at inference time.
This work formalizes target-mode coverage as target-mode coverage, the coverage of a predefined set of semantically specified modes, and proposes multi-axis max@K, a group-based reinforcement learning objective for improving it in diffusion-based T2I models.
Ku Onoda, Paavo Parmas, Hiroki Furuta et al.· arXiv.org· 3 citations
The Regional Self-Attention Mechanism is proposed, which strictly confines the self-attention operation to the independent regions defined by the target bounding boxes, enabling the model to generate specific-category targets in a targeted manner during the denoising process.
Haoshu Zhao, Xi Zhang· International Conference on...· 0 citations
A training-free Dual-path Attention Modulation (DAR) framework that decouples semantic edits while preserving source image structure is proposed and Adaptive Self-Attention (ASA) and Adaptive Cross-Attention (ACA) modules that dynamically regulate attention replacement are introduced.
Tong Cui, Jie Yang, Kai-Ru Li et al.· International Conference on...· 0 citations
Text-to-image diffusion models have achieved remarkable success in generating high-quality images from a given text prompt. Subject-driven generation aims to synthesize customized images to mimic the appearance of subjects in given reference images within different visual contexts specified by the text prompts. The cen...
Yushun Tang, Weiming Chen, Siyi Liu et al.· IEEE transactions on multime...· 0 citations
Subject-driven image personalization---generating new images that preserve the identity of one or several reference subjects in novel scenes---is a foundational capability for modern visual content creation. It is currently dominated by generalized methods that fine-tune a pretrained multimodal diffusion transformer (M...
Jihun Park, Kyoungmin Lee, Jongmin Gim et al.· 0 citations
The paradigms of the generative adversarial networks (GANs), variational autoencoders (VAEs), transformer-based designs, and diffusion models, with the last one representing the state of the art in image generation models are discussed.
