ACDG-VTON: Accurate and Contained Diffusion Generation for Virtual Try-On
This addresses garment identity preservation in virtual try-on for fashion and e-commerce applications, representing an incremental improvement over existing diffusion methods.
The paper tackles the problem of diffusion-based virtual try-on methods struggling to preserve garment identities by proposing a unique training scheme that uses a control image to limit diffusion scope, resulting in accurate preservation of garment details and support for layering, styling, and shoe try-on in a single inference cycle.
Virtual Try-on (VTON) involves generating images of a person wearing selected garments. Diffusion-based methods, in particular, can create high-quality images, but they struggle to maintain the identities of the input garments. We identified this problem stems from the specifics in the training formulation for diffusion. To address this, we propose a unique training scheme that limits the scope in which diffusion is trained. We use a control image that perfectly aligns with the target image during training. In turn, this accurately preserves garment details during inference. We demonstrate our method not only effectively conserves garment details but also allows for layering, styling, and shoe try-on. Our method runs multi-garment try-on in a single inference cycle and can support high-quality zoomed-in generations without training in higher resolutions. Finally, we show our method surpasses prior methods in accuracy and quality.