Abstract:
The fashion industry is facing a significant challenge with the rise of fast fashion, leading to wardrobe
clutter and environmental concerns. Young adults, in particular, struggle with managing their
wardrobes efficiently, often purchasing duplicate items and failing to utilize their clothing to its fullest
potential. The lack of a comprehensive tool for wardrobe management exacerbates the problem,
resulting in wasteful consumption patterns and a disorganized approach to personal style.
To address these challenges, Wardrobe Whiz was developed as a virtual wardrobe assistant
leveraging cutting-edge technologies. Utilizing the Segformer model for semantic segmentation, the
platform accurately identifies and categorizes clothing items from user-uploaded images, enabling
precise inventory management. The Gemini model is employed for feature extraction, capturing
intricate fashion details and facilitating sophisticated outfit recommendations based on style
compatibility. The recommendation engine, powered by a combination of collaborative filtering and
content-based filtering, provides personalized outfit suggestions tailored to occasions and weather
conditions, enhancing the user's decision-making process and promoting sustainable fashion choices.
Testing and evaluation of Wardrobe Whiz's functionalities have shown promising results, with the
Segformer model demonstrating high accuracy in segmenting clothing items. The recommendation
engine's effectiveness in providing relevant and diverse outfit suggestions has also been noteworthy.
These results underscore the potential of Wardrobe Whiz as a comprehensive solution for modern
wardrobe management, combining technological innovation with a commitment to sustainability.
