An Application to Generate Style Guided Compatible Outfit
This addresses fashion recommendation for consumers by focusing on style-guided outfit generation, which is incremental as it builds on existing compatibility learning work.
The paper tackles the problem of generating fashion outfits guided by styles or themes, using a novel style encoder network to create compatible outfits based on an anchor item and styles, with results demonstrated through experiments and an API.
Fashion recommendation has witnessed a phenomenal growth of research, particularly in the domains of shop-the-look, contextaware outfit creation, personalizing outfit creation etc. Majority of the work in this area focuses on better understanding of the notion of complimentary relationship between lifestyle items. Quite recently, some works have realised that style plays a vital role in fashion, especially in the understanding of compatibility learning and outfit creation. In this paper, we would like to present the end-to-end design of a methodology in which we aim to generate outfits guided by styles or themes using a novel style encoder network. We present an extensive analysis of different aspects of our method through various experiments. We also provide a demonstration api to showcase the ability of our work in generating outfits based on an anchor item and styles.