Image-Derived Friend Avatars for AR Fashion Try-On
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality (AR) systems require users to navigate between multiple interfaces to try on virtual items, wasting time and resources, and there is no efficient way for users to visualize how fashion items would look on their friends without explicitly requesting their avatar.
Innovation Solution
A machine learning model generates an avatar based on images of a friend, allowing users to apply AR fashion items to the avatar for a realistic try-on experience without transmitting the friend's avatar, reducing resource consumption and improving user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users navigate between multiple interfaces to try on virtual items, then the AR try-on functionality is achieved, but time and resources are wasted
Solution Approach 1:
The patent combines the product browsing interface and AR try-on interface into a single unified interface. Users can view products and activate AR try-on experiences within the same screen without navigating away, eliminating the need to switch between multiple interfaces and reducing navigation time while maintaining full AR functionality.
Solution Approach 2:
The unified interface serves multiple functions simultaneously: it displays product information, provides product selection capabilities, and enables AR try-on experiences all in one view. This multi-functional design allows users to access AR try-on features directly from the product browsing context without requiring separate navigation steps.
2Measurement precision
If users explicitly request friend's avatar for try-on, then accurate friend visualization is achieved, but user privacy and convenience are reduced
Solution Approach 1:
The system creates a digital copy or representation of the friend's appearance using available data (such as photos from social media profiles or messaging apps) without requiring the friend's actual avatar file or explicit permission. This copy enables accurate visualization of how fashion items would look on the friend while maintaining privacy and convenience, as no direct avatar transmission or explicit request is needed.
Solution Approach 2:
The system introduces an intermediary processing layer that generates friend appearance representations from indirect data sources rather than requiring direct access to the friend's avatar. This intermediary mechanism allows the system to infer friend appearance characteristics from available information, enabling accurate try-on visualization while avoiding the need for explicit friend authorization or direct avatar file access.
3Measurement precision
If avatars are transmitted between users for try-on, then accurate friend visualization is achieved, but resource consumption increases
Solution Approach 1:
The system extracts only the essential appearance features needed for accurate friend visualization (such as body shape, size, and key physical characteristics) from source images, rather than transmitting complete avatar files. This extraction approach maintains visualization accuracy while significantly reducing data transmission requirements and associated energy consumption.
Solution Approach 2:
Instead of transmitting original avatar data between users, the system creates lightweight digital copies or synthesized representations of friend appearance locally on the user's device. These copied representations contain sufficient detail for accurate try-on visualization but require minimal storage and transmission resources, thereby reducing energy consumption while maintaining visualization fidelity.
Data Source
AI summary
Aspects of the present disclosure involve a system for providing an AR try-on experience for a friend. The system accesses a plurality of images that depict one or more persons. The system receives input that identifies a given person of the one or more persons who is depicted in an individual image of the plurality of images. The system extracts features of the given person depicted in the individual image. The system applies a machine learning model to the extracted features of the given person to generate an avatar that resembles the given person. The system applies one or more augmented reality (AR) fashion items to the avatar to generate an image that resembles the given person wearing the one or more AR fashion items.