Final Year Project

Project Title

Personalized Outfit Recommendation System

Description

One of the biggest problems in fashion retail is product curation. Retailers have to spend a large amount of time to come up with different combinations of their products that would as a whole, go well as an outfit, and even then, the options aren’t really personalized. A customer buys a new shirt, brings it home, and hangs it up, only to find that the shirt stays in their closet for weeks because they’re not sure what to pair it with. This also means a loss in conversion rates and potential revenue at the side of the retailer. We see a business opportunity in this problem, and so the idea behind the project is to solve it by addressing the key issue, product curation, by providing expert recommendations across different clothing items to the end-consumer at the point of sale or as a standalone service. Our tool allows shoppers to visually search the catalog of e-commerce stores by uploading pictures of outfits they like or taking a photo with their phone’s camera. Using Computer Vision and Machine Learning, identical and/or visually similar items from the store are shown. This allows shoppers to quickly and conveniently shop for items they see on social media, significantly increasing the conversion rate. (Group Project)