双通道注意力引导用于扩散变换器中无需训练的图像编辑控制
Dual-Channel Attention Guidance for Training-Free Image Editing Control in Diffusion Transformers
AI总结：
本文提出双通道注意力引导方法，通过同时操控键通道和值通道实现无需训练的图像编辑控制，显著提升编辑保真度。
摘要：
在基于扩散变换器（DiT）架构的扩散图像编辑模型中，无需训练的编辑强度控制是一个关键要求。现有注意力操纵方法仅专注于键空间以调节注意力路由，忽略了值空间——这决定了特征聚合。在本文中，我们首先揭示了DiT多模态注意力层中的键和值投影表现出明显的偏置-delta结构，其中标记嵌入紧密聚集在特定层的偏置向量周围。基于这一观察，我们提出了双通道注意力引导（DCAG），这是一个无需训练的框架，同时操纵键通道（控制注意力的位置）和值通道（控制聚合的内容）。我们提供了一个理论分析，表明键通道通过非线性softmax函数操作，作为粗控制旋钮，而值通道通过线性加权求和操作，作为细粒度的补充。共同，二维参数空间（δ_k，δ_v）比任何单通道方法都能实现更精确的编辑保真度权衡。在PIE-Bench基准（700张图像，10个编辑类别）上的大量实验表明，DCAG在所有保真度指标上均优于仅键引导的方法，最显著的改进出现在局部编辑任务中，如物体删除（LPIPS减少4.9%）和物体添加（LPIPS减少3.2%）。\n
英文摘要：
Training-free control over editing intensity is a critical requirement for diffusion-based image editing models built on the Diffusion Transformer (DiT) architecture. Existing attention manipulation methods focus exclusively on the Key space to modulate attention routing, leaving the Value space -- which governs feature aggregation -- entirely unexploited. In this paper, we first reveal that both Key and Value projections in DiT's multi-modal attention layers exhibit a pronounced bias-delta structure, where token embeddings cluster tightly around a layer-specific bias vector. Building on this observation, we propose Dual-Channel Attention Guidance (DCAG), a training-free framework that simultaneously manipulates both the Key channel (controlling where to attend) and the Value channel (controlling what to aggregate). We provide a theoretical analysis showing that the Key channel operates through the nonlinear softmax function, acting as a coarse control knob, while the Value channel operates through linear weighted summation, serving as a fine-grained complement. Together, the two-dimensional parameter space $(δ_k, δ_v)$ enables more precise editing-fidelity trade-offs than any single-channel method. Extensive experiments on the PIE-Bench benchmark (700 images, 10 editing categories) demonstrate that DCAG consistently outperforms Key-only guidance across all fidelity metrics, with the most significant improvements observed in localized editing tasks such as object deletion (4.9% LPIPS reduction) and object addition (3.2% LPIPS reduction).
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
