Hi, magic closet, tell me what to wear!

Abstract

In this paper, we aim at a practical system, magic closet, for automatic occasion-oriented clothing recommendation. Given a user-input occasion, e.g., wedding, shopping or dating, magic closet intelligently suggests the most suitable clothing from the user's own clothing photo album, or automatically pairs the user-specified reference clothing (upperbody or lower-body) with the most suitable one from online shops.Two key criteria are explicitly considered for the magic closet system. One criterion is to wear properly, e.g., compared to suit pants, it is more decent to wear a cocktail dress for a banquet occasion. The other criterion is to wear aesthetically, e.g., a red T-shirt matches better white pants than green pants. To narrow the semantic gap between the low-level features of clothing and the high-level occasion categories, we adopt middle-level clothing attributes (e.g., clothing category, color, pattern) as a bridge. More specifically, the clothing attributes are treated as latent variables in our proposed latent Support Vector Machine (SVM) based recommendation model. The wearing properly criterion is described in the model through a feature-occasion potential and an attribute-occasion potential, while the wearing aesthetically criterion is expressed by an attribute-attribute potential. To learn a generalize-well model and comprehensively evaluate it, we collect a large clothing What-to-Wear (WoW) dataset, and thoroughly annotate the whole dataset with 7 multi-value clothing attributes and 10 occasion categories via Amazon Mechanic Turk. Extensive experiments on the WoW dataset demonstrate the effectiveness of the magic closet system for both occasion-oriented clothing recommendation and pairing. Figure 1: Scenario illustration of magic closet. A user specifies an occasion, and the most suitable clothing are recommended from the user's mobile photo album. For going on a date, the recommended clothing are sweet, have bright color and skirts are preferred. For better view, please refer to the original color pdf file.

Outfit Recommendation System Based on Deep Learning

“…Secondly, some outfit related applications such as identifying people based on their outfits then to predict occupation [11], finding outfit similarity [12], recommending good outfits to users [13], and predicting the existed outfit style [14] are the works related to our goal. However, most of the approaches here use annotated data, which is not suitable for general situations.…”

“…[13], both of them use manually annotated data like above. But there are other systems use unannotated data like [5] uses data from Polyvore.com, just the same as ours, to recommend items to users according to their personal preference.…”

Runway to Realway: Visual Analysis of Fashion

“…Two key differences are the direction of the matching and the focus on outfits rather that clothing items. Our trend analyses could also enable new fashion applications, for example, by improving market analysis of fashion brands [13], interactive search [14] or recommendation [12,17] of trending styles.…”

Regularized deep learning for fashion and social media data analysis

“…These studies include clothing parsing [3], fashion style classification [5,16], clothing retrieval [11,12], clothing recommendation [86] and so on. Our work is mainly related to clothing retrieval and clothing recommendation.…”

“…Accessories recommendation, or namely, clothing pairing getting hotter in recent years [92,86,91]. Wata et al [92] Simo-Serra et al aims at estimating the fashionability of pairing two accessories [93].…”