Deep Learning for Automated Tagging of Fashion Images

Patricia Gutierrez, Pierre-Antoine Sondag, Petar Butkovic, Mauro Lacy, Jordi Berges, Felipe Bertrand, Arne Knudson

Abstract: We present 9 deep learning classifiers to predict Fashion attributes in 4 different categories: apparel (dresses and tops), shoes, watches and luggages. Our prediction system hosts several classifiers working at scale to populate a catalogue of millions of products. We provide details of our models as well as the challenges involved in predicting Fashion attributes in a relatively homogeneous problem space.