用真人穿搭图生成高保真服装平铺图，支持个性化推荐与虚拟试穿。
TryOffAnyone: Tiled Cloth Generation from a Dressed Person
- 基于微调的Stable Diffusion，单阶段架构结合服装掩码精准分割目标服饰。
- 在VITON-HD数据集上实现当前最佳性能，全身与半身输入均生成高质量结果。
- 轻量化设计减少计算开销，适合实际应用，开源代码与模型可复现。
时尚产业正越来越多地利用计算机视觉与深度学习技术提升在线购物体验与运营效率。本文针对从模特穿着照片中生成高保真平铺服装图像这一挑战，提出一种新方法。受潜在扩散模型（LDM）在图像到图像翻译中的成功启发，我们采用微调的StableDiffusion模型，设计了一种简化的单阶段网络结构，通过特定服装掩码有效隔离并处理目标衣物。通过选择性训练Transformer块并移除冗余交叉注意力层，显著降低计算复杂度，同时在VITON-HD等基准数据集上达到当前最优表现。实验表明，该方法对全身和半身输入均能生成高质量的平铺服装图像。代码与模型已公开于：https://github.com/ixarchakos/try-off-anyone。
原文摘要 · Abstract (English)
The fashion industry is increasingly leveraging computer vision and deep learning technologies to enhance online shopping experiences and operational efficiencies. In this paper, we address the challenge of generating high-fidelity tiled garment images essential for personalized recommendations, outfit composition, and virtual try-on systems from photos of garments worn by models. Inspired by the success of Latent Diffusion Models (LDMs) in image-to-image translation, we propose a novel approach utilizing a fine-tuned StableDiffusion model. Our method features a streamlined single-stage network design, which integrates garmentspecific masks to isolate and process target clothing items effectively. By simplifying the network architecture through selective training of transformer blocks and removing unnecessary crossattention layers, we significantly reduce computational complexity while achieving state-of-the-art performance on benchmark datasets like VITON-HD. Experimental results demonstrate the effectiveness of our approach in producing high-quality tiled garment images for both full-body and half-body inputs. Code and model are available at: https://github.com/ixarchakos/try-off-anyone
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。