US20260099961
2026-04-09
Physics
G06T11/00
Diffusion Model Virtual Try-On Experience introduces methods and systems for creating an extended reality (XR) try-on experience using diffusion models. The system processes a real-world object image and a target fashion item image to generate a warped image. This image is then analyzed using a generative machine learning model to produce an artificial image that depicts the real-world object wearing the fashion item. This approach enhances the virtual try-on experience by providing realistic and high-quality images with minimal user input.
Augmented reality (AR) and extended reality (XR) systems blend virtual and real-world environments, allowing users to interact with virtual objects in a realistic manner. These technologies are used in various applications, such as video games and communication platforms. However, creating high-quality images for these systems often requires expensive equipment and significant effort, which can limit their accessibility and effectiveness.
The disclosed system improves image generation efficiency by automating the creation of photorealistic images of real-world objects in simulated scenes. It uses a diffusion model to replace pixels of a real-world object with those of a target fashion item, generating a warped image. Segmentation maps are created for incomplete image portions, and a generative machine learning model fills these gaps, producing a complete artificial image. This process reduces the time and resources needed to create high-quality images for XR experiences.
The system operates within a networked computing environment, involving multiple user systems and interaction clients. These clients communicate over a network with interaction server systems and third-party servers, exchanging data such as text, audio, and video. The interaction server system provides server-side functionality and processes data from interaction clients, supporting various services and operations necessary for the XR try-on experience.
Interaction clients can access features of external resources, such as linked applications or applets, through their interfaces. These external resources, which may be provided by third parties or the interaction client provider, offer additional functionality and are often implemented using markup-language documents. This integration allows users to seamlessly access and utilize a range of applications and services within the XR environment.