提出解耦服装条件与去噪过程的新方法，用单网络实现双网络级效果。
Rethinking Garment Conditioning in Diffusion-based Virtual Try-On: Decouple, Don't Denoise
- 将服装条件与去噪过程分离，避免信息干扰
- 在相同参数量下超越现有单网络方法，媲美双网络性能
- 无需修改架构，适合电商和时尚领域的图像合成应用
虚拟试衣（VTON）旨在生成人物穿戴目标服装的逼真图像，广泛应用于电商与时尚领域。基于扩散模型的双UNet方法虽表现优异，但需额外网络进行服装条件建模，参数量翻倍。空间拼接方案虽为单网络设计，但无论是基于UNet还是DiT的实现均表明全微调无效，社区普遍采用仅训练注意力层的方式。本文首次对双UNet参考网络行为进行可视化分析，发现服装条件必须与去噪过程解耦。空间拼接因将服装嵌入去噪目标，导致引导泄漏、梯度竞争与训练测试不一致三大冲突。据此提出三项设计原则，并以纯配方形式应用于标准架构，无需修改结构。所提DeCo-VTON模型（860M参数）达到单网络最佳性能，以一半参数量媲美双网络最优结果，且在人类评估中更受青睐。
原文摘要 · Abstract (English)
Virtual Try-On (VTON) synthesizes realistic images of a person wearing a target garment, with broad applications in e-commerce and fashion. Diffusion-based dual-UNet methods achieve strong results but double the parameters by dedicating a separate network to garment conditioning. Spatial concatenation offers a simpler single-network alternative, yet both UNet- and DiT-based instantiations report that full fine-tuning is ineffective, and the community has settled for attention-only training. We ask: why does full fine-tuning fail, and can this be resolved? Through what is, to our knowledge, the first visualization study of dual-UNet reference network behavior, we identify a unifying insight: garment conditioning must be decoupled from the denoising process. Spatial concatenation violates this by embedding the garment within the denoising target, causing three conflicts: guidance leakage, gradient competition, and train-test discrepancy. We derive three design principles to restore this decoupling and implement them as a pure recipe atop a standard architecture with no modification. The resulting model, DeCo-VTON (860M params), achieves single-network state of the art, matching the dual-UNet state of the art at half the cost while being preferred in human evaluation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。