Clothing Visual Search

LANGUAGE Python, Java

TIME Capstone Project, Septmeber 2019 - April 2020


An android app which runs a neural network to return pictures of clothing items similar to a user-provided input.

The growing ubiquity of e-commerce in the fashion industry has further emphasised the inherent drawbacks with the traditional text-based search that drives many large e-commerce conglomerates. Whether due to semantic ambiguity in text queries or misspelled words for example, text-based search leaves much to be desired when it comes to a user attempting to describe a clothing item using words.

As a result, our team designed a visual search application. The goal of our project is twofold: to develop a deep learning framework that returns pictures of clothing items similar to a user-provided image (Goal ML) and to develop a mobile application to showcase our design by simulating an arbitrary fashion retailer (Goal App.).