We love it for keeping in touch with friends and family and catching up on all the latest news, but we might find ourselves turning to Facebook to help us up our fashion game too. The social network is experimenting on a new artificial intelligence technology that may help users make minimal edits to their outfits for an instant and easy fashion fix, a service that could be described as an online personal stylist.
With its ability to recognise garments, AI system Fashion++ offers subtle options to make an ensemble look better. By using it as a guide to rock a more fashionable look, users can benefit by making small and simple adjustments that can create a big difference, whether it’s rolling up the sleeves, ditching an accessory or swapping one shirt for another. What’s more, Fashion++, which uses a deep image-generation neural network to make the suggestions, can help users save money as the simple tweaks it recommends can help step up that fashion A-game without having to buy anything new.
"Whereas previous work in this area has explored ways to recommend an entirely new outfit or to identify garments that are similar to one another, Fashion++ instead aims to suggest subtle alterations to an existing outfit that will make it more stylish," Facebook noted while describing the programme.
The research project is an example of how AI can be used in a creative and practical application. And it seems to be a success as when the research team asked human evaluators to rate Fashion++’s advice, the participants preferred its suggestions.
“The system uses a discriminative fashionability classifier that is trained on thousands of publicly available images of outfits that have been judged to be stylish. These serve as ground truth examples of fashionable outfits, and unfashionable examples are then bootstrapped by swapping garments on the fashionable examples with their least similar counterparts,” Facebook AI state in the official breakdown of how it works.
“Once the classifier is trained, our system gradually updates the outfit in order to make it more fashionable. An image-generation neural network renders the newly adjusted look, using a variational auto-encoder to generate the silhouette and a conditional generative adversarial network (cGAN) to generate the colour and pattern. The latent encodings learned by this generator are further used to identify which garments in its inventory will best achieve the style.”
