Smart Wardrobe : An Apparel Recommendation System

Abstract

Recommendation systems are a category of information filtering systems that represents user preferences. This project presents an Android based outfit recommendation system which suggests items for both males and females based on their preferences. The user can upload query image and request for clothing items with similar color or pattern. User can further enhance the search by manually selecting outfit categories such as jackets, sweaters,T-shirts etc. The system uses HSI color histogram matching for retrieving similar colored items and applies Gabor filtering for texture analysis. Results suggest that our system achieves promising classification rates in retrieving relevant clothing items.