Clothing Matching for Visually Impaired Persons.

Area of Science:

• Computer Vision
• Assistive Technology
• Human-Computer Interaction

Background:

• Matching clothes is a significant challenge for individuals with visual impairments.
• Existing solutions often lack the ability to handle complex patterns or color deficiencies.
• There is a need for an automated system to aid in clothing selection for the visually impaired.

Purpose of the Study:

• To develop and evaluate a proof-of-concept system for automated clothing matching.
• To assist blind and color-deficient individuals in selecting matching garments.
• To create a robust system that handles variations in clothing and lighting.

Main Methods:

• A system combining a camera, computer vision algorithms for pattern and color matching, speech commands, and audio feedback was developed.
• The system processes clothing images to identify and match colors and patterns.
• Databases of clothing images with varying patterns, colors, and under different lighting/rotation conditions were collected for evaluation.

Main Results:

• The proposed system demonstrates robustness against variations in illumination, clothing rotation, and wrinkling.
• It effectively matches both simple (no pattern, deficient color) and complex (multiple colors, intricate patterns) clothing items.
• Evaluation results confirm the system's effectiveness in clothing matching for visually impaired users.

Conclusions:

• The developed clothing matching system offers a viable technological solution for a common challenge faced by blind and color-deficient individuals.
• The system's robustness and effectiveness in diverse conditions highlight its potential for practical application.
• This proof-of-concept system paves the way for enhanced independence and confidence in clothing selection for users with visual impairments.