Avatar Region Classification for Realistic Virtual Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in virtually trying on clothing using large format display devices, as they cannot accurately predict how clothing will look when worn, and the service is not well-suited to changing seasons and surroundings.
Innovation Solution
An electronic device that captures a user's image, sets an avatar region, generates an avatar based on body information, and combines it with the user's image to create a composite display, allowing for a realistic virtual fitting service by classifying the display into a mirror region and an avatar region.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a user views clothing on an avatar or model displayed on an LFD device, then the user can see how the clothing looks on others, but the user cannot accurately predict how the clothing will look when actually worn
Solution Approach 1:
The patent creates a virtual copy (avatar) of the user's actual body by capturing images with a camera and generating a three-dimensional avatar model. This copy allows the user to see how clothing would look on their own body shape and features, providing accurate prediction while maintaining the convenience of virtual fitting.
Solution Approach 2:
The patent transitions from two-dimensional images displayed on traditional screens to three-dimensional avatars that can be viewed from multiple angles. This dimensional enhancement allows users to rotate and examine the avatar with clothing from different perspectives, significantly improving prediction accuracy of how the clothing will look when worn.
2Adaptability or versatility
If traditional virtual fitting services are used, then clothing can be viewed on a model, but the service is not well-suited to changing seasons and surroundings
Solution Approach 1:
The patent implements dynamic backgrounds that can change according to different seasons, times of day, and environments. Instead of static model images, the system allows users to place their three-dimensional avatar in various virtual settings (e.g., beach for summer, snowy landscape for winter), making the fitting service adaptable to changing seasons and surroundings while maintaining reliable suitability assessment.
Solution Approach 2:
The patent creates a universal virtual fitting platform that works across different seasons, locations, and clothing types. The three-dimensional avatar can be dressed in any clothing item and placed in any virtual environment, making the service universally applicable regardless of seasonal or environmental changes.
3Ease of operation
If an LFD device is used for virtual fitting, then large format display provides visual information, but the display does not provide a natural and realistic virtual fitting service
Solution Approach 1:
The patent changes the key parameter of visualization from two-dimensional flat images to three-dimensional models with depth, volume, and spatial orientation. This parameter change allows the avatar to be rotated 360 degrees and viewed from any angle, providing a natural and realistic representation of how clothing drapes and fits on the body, while still utilizing the large format display for comprehensive visual information.
Data Source
AI summary
A method and an electronic device for composing an image are provided. An electronic device includes a display configured to display an image of a user photographed by a camera; an input component configured to receive a user input; a communicator configured to facilitate a communication with an external server; and a processor configured to display the image of the user on the display, set an avatar region within the image of the user based on the received user input, generate an avatar to be displayed on the avatar region, and control the display to combine the avatar with the image of the user and to display the resulting composite image.


