Clothing Registration and Recommendation Using Image-Based Virtual Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtual fitting systems are limited in recommending suitable clothing as they require pre-stored clothing information and cannot account for body changes or suggest clothing for unsold items, and they lack integration with home appliances for clothing care, leading to inefficient clothing selection and care practices.
Innovation Solution
A home appliance system that uses image recognition and analysis to compute clothing information, recommend suitable clothing based on weather, schedule, and body changes, and integrates with a virtual fitting service to suggest purchases and guide clothing care through a connected online system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional virtual fitting systems use pre-stored clothing information for rendering, then virtual fitting can be performed on specific clothing provided by sellers, but the system cannot perform virtual fitting on clothing whose information is not provided or on user's stored clothing
Solution Approach 1:
The system performs preliminary actions by having users upload images of their stored clothing in advance. These images are then analyzed and stored as clothing information, enabling future virtual fitting operations on user's own clothing without requiring pre-existing structured clothing data from sellers.
Solution Approach 2:
The system creates copies of clothing information by analyzing images of physical clothing items. Instead of requiring original structured clothing data, the system generates digital representations (copies) of clothing from photographs, allowing virtual fitting on any clothing item that can be photographed.
2Adaptability or versatility
If users purchase multiple clothing items from different sellers, then users have more clothing choices, but users cannot perform virtual fitting on these clothing items as sellers do not provide clothing information
Solution Approach 1:
The system enables self-service by allowing users to independently upload and register their own clothing images. Users can create their personal clothing database without relying on sellers to provide structured clothing information, thus enabling virtual fitting on any clothing they own regardless of seller participation.
3Productivity
If conventional systems recommend clothing based on pre-stored information, then recommendations can be made for available clothing, but the system cannot account for body changes or suggest clothing for unsold items
Solution Approach 1:
The system implements feedback mechanisms by repeatedly imaging the user's body over time and detecting body changes. This continuous feedback loop allows the system to update the user's body model and adjust clothing recommendations accordingly, ensuring recommendations remain accurate as the user's body changes.
4Device complexity
If virtual fitting systems require pre-stored clothing information, then the system structure remains simple, but the system lacks integration with home appliances for clothing care
Solution Approach 1:
The system achieves multi-functionality by integrating clothing recommendation capabilities with home appliances such as washing machines and dryers. The same system that recommends clothing can now also provide care guidance for clothing items, allowing a single system to serve multiple functions including recommendation, care instruction, and appliance control.
Data Source
AI summary
The present invention relates to an online system and a control method therefor, the online system comprising clothing devices which may receive physical information of a user, notify of a change in the physical information, register clothing information held by the user, put together the schedule of the user, weather, preferences, the trend in the area where the user is located, and information on the clothing being sold, and recommend customized clothing to the user.