Video virtual try-on preserves garment textures in high resolution
TexTailor: Texture-Preserving Video Virtual Try-On via Adaptive Garment Conditioning
Computer Vision and Pattern Recognition
Summary
Virtual try-on technology lets people see how clothes look on them in videos. Existing methods struggle with high-resolution videos and often lose clothing details or smoothness over time. The authors created TexTailor, which uses a new video diffusion model to keep clothing textures clear and consistent across frames. Their approach also better matches clothing with the video, leading to higher quality, more detailed try-on videos.
What this means in practice
- •For fashion e-commerce teams: Create high-quality virtual try-on videos that preserve fine fabric details for online clothing stores.$Commercial implications: Allows companies to offer better virtual fitting experiences, improving digital fashion shopping and customer engagement.
- •For video content creators: Produce virtual clothing videos with improved detail and smooth appearance over time for fashion shows or social media.
Authors
Zijing Qin, Jun Zhou, Ruicheng Zhang, Jiaqi Hou, Zunnan Xu, Ronghui Li, Zhenyu Xie, Xiu Li
Abstract
Video virtual try-on has attracted increasing attention due to its broad potential in digital fashion and intelligent e-commerce. However, existing methods primarily focus on low-resolution settings and still face substantial challenges when extended to high-resolution scenarios. These limitations can be attributed to two main factors: (1) the insufficient utilization of rich garment reference information, and (2) the lack of explicit positional modeling between garment and video representations during cross-modal interaction, which weakens fine-grained local correspondence. To address these issues, we propose TexTailor, a high-fidelity video virtual try-on framework built upon a pretrained video Diffusion Transformer. Specifically, we introduce a timestep-adaptive modulation mechanism to dynamically adjust garment visual representations throughout denoising. We further develop a frame-aligned positional encoding strategy to strengthen garment-to-video correspondence, together with a multi-source injection design that reduces interference among heterogeneous conditions. Extensive experiments on multiple video virtual try-on benchmarks, including the high-resolution Eevee dataset, demonstrate that TexTailor achieves competitive performance in garment detail preservation, temporal consistency, and overall video quality.