Fashion Recommender

This project presents a novel framework for fashion recommendation that is driven by data, visually related, and simple effective recommendation systems for generating fashion product images. The proposed approach uses a two-stage phase, initially extracting features of the image using a CNN classifier, and then recommending appealing outfits using a multi-input CNN and K-Means clustering. The system takes separate articles of clothing as image inputs and recommends what it considers to be appealing outfits. The final user interface is created with a Dataiku application. The project uses over 100,000 images from the Polyvore dataset and achieves promising results. The system can be extended to recommend articles of clothing to users to enhance their existing wardrobe or dress store mannequins.

‍