US20240394991
2024-11-28
Physics
G06T19/20
The virtual fitting system utilizes advanced technology to simulate how clothing fits on a user's avatar in a realistic manner. It employs an artificial intelligence server to process 3D data of the user's avatar and selected clothing, using neural network operations to generate virtual fitting data. A rendering server then creates a visual representation of this data, which is transmitted to the user via an image transmission server. A database stores all relevant 3D data and material parameters, ensuring that the virtual fitting is both accurate and personalized.
Traditional virtual fitting services relied on the hardware capabilities of user devices, often resulting in 2D images that lacked realism and failed to reflect individual body types accurately. These services required users to operate separate applications, adding unnecessary complexity to the online shopping experience. The new system overcomes these limitations by performing 3D operations on external servers, independent of the user's device performance.
The proposed solution offers several advantages over conventional methods. It can produce realistic 3D fitting images that account for various user poses and clothing material dynamics. This system integrates directly with online shopping applications, removing the need for additional virtual fitting apps and streamlining the shopping process. The AI server is trained with physics-based losses to ensure high fidelity in clothing simulation.
The virtual fitting device operates through several key components: an AI server for processing avatar and clothing data, a rendering server for generating visual outputs, and an image transmission server for delivering these outputs to users. The system's database holds detailed 3D models and material parameters, which are essential for accurate simulations. The AI server uses neural networks trained with specific loss functions to enhance realism in clothing behavior.
Users can access the virtual fitting service directly from their preferred online shopping applications, where they can select clothing items for virtual trials. The service delivers results within the shopping app interface, such as through pop-up windows, providing a seamless and convenient experience. This integration eliminates the need for separate applications, making it easier for users to visualize how different garments will fit them before making a purchase.