AI-Based Garment Classification for Relaxed and Crumpled Forms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing garment classification technologies require garments to be neatly presented and in a flat form to be classified, failing to accurately classify garments in relaxed or arbitrary forms such as crumpled or folded states.
Innovation Solution
A customizable garment sorting system utilizing AI technology with modules for type, fabric structure, fabric material, and color classification, capable of classifying garments in arbitrary forms through deep learning algorithms and hyperspectral imaging, integrated with a robotic arm for automated sorting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If garments are required to be neatly presented in flat form for classification, then classification accuracy is improved, but operational complexity and time consumption increase
Solution Approach 1:
The patent replaces manual mechanical inspection and flat-laying requirements with an automated image recognition system using deep learning algorithms. The system captures images of garments in their natural relaxed states and automatically classifies them by type, fabric structure, material, and color, eliminating the need for manual preparation while maintaining high classification accuracy.
Solution Approach 2:
The patent changes the input parameter from requiring garments to be in a specific flat state to accepting garments in any relaxed state. The deep learning model is trained to recognize garment features regardless of their specific configuration, allowing classification based on image data alone without mechanical preparation.
2Measurement precision
If garments are required to be neatly presented in flat form for classification, then classification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent replaces complex mechanical systems designed to flatten and position garments with a simpler optical system consisting of cameras and lighting. The classification function is transferred from mechanical manipulation to software-based image recognition using deep learning algorithms.
Solution Approach 2:
The patent extracts the classification function from the physical manipulation of garments and transfers it to a digital image analysis system. By separating the classification task from the physical state requirements, the system achieves accuracy without the complexity of mechanical preparation devices.
3Measurement precision
If garments are required to be neatly presented in flat form for classification, then classification accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The patent replaces manual garment preparation and inspection with an automated imaging and classification system. Operators simply need to place garments in a pile or on a surface, and the system automatically captures images and performs classification, dramatically improving ease of operation while maintaining accuracy.
Solution Approach 2:
The system performs self-service by automatically capturing images, processing them through deep learning algorithms, and generating classification results without requiring manual intervention for garment preparation or inspection. The system serves itself by handling the entire classification process autonomously.
Data Source
AI summary
A garment classifying and sorting system and method which sort and classify garments even when the garments are in a relaxed form and arbitrary form, such as when the garment is crumpled, wrinkled or folded.