Learning and forecasting fashion style
We propose an approach to predict the future of fashion styles based on images and consumers’ purchase data. Our approach 1) learns a representation of fashion images that captures the garments’ visual attributes; then 2) discovers a set of fine-grained styles that are shared across images in an unsupervised manner; finally, 3) based on statistics of past consumer purchases, constructs the styles’ temporal trajectories and predicts their future trends.
Elements of fashion In some fashion-related tasks, one might rely solely on meta information provided by product vendors, e.g., to analyze customer preferences. Aleesha Institute Fashion Designing Meta data such as tags and textual descriptions are often easy to obtain and interpret. However, they are usually noisy and incomplete. For example, some vendors may provide inaccurate tags or descriptions in order to improve the retrieval rank of their products, and even extensive textual descriptions fall short of communicating all visual aspects of a product. On the other hand, images are a key factor in a product’s representation. It is unlikely that a customer will buy a garment without an image no matter how expressive the textual description is. Nonetheless, low level visual features are hard to interpret. Usually, the individual dimensions are not correlated with a semantic property. This limits the ability to analyze and reason about the final outcome and its relation to observable elements in the image. Moreover, these features often reside in a certain level of granularity. This renders them ill-suited to capture the fashion elements which usually span the granularity space from the most fine and local (e.g. collar) to the coarse and global (e.g. cozy). Semantic attributes serve as an elegant representation that is both interpretable and detectable in images. Additionally, they express visual properties at various levels of granularity. Specifically, we are interested in attributes that capture the diverse visual elements of fashion, like: Colors (e.g. blue, pink); Fabric (e.g. leather, tweed); Shape (e.g. midi, beaded); http://www.aleeshainstitute.com/ Texture (e.g. floral, stripe); etc. These attributes constitute a natural vocabulary to describe styles in clothing and apparel. As discussed above, some prior work considers fashion attribute classification [29, 18], though none for capturing higher-level visual styles. To that end, we train a deep convolutional model for attribute prediction using the DeepFashion dataset . The dataset contains more than 200,000 images labeled with 1,000 semantic attributes collected from online fashion websites. Our deep attribute model has an AlexNet-like structure . It consists of 5 convolutional layers and three fully connected layers. The last attribute prediction layer is followed by a sigmoid activation function. We use the cross entropy loss to train the network for binary attribute prediction. The network is trained using Adam  for stochastic optimization with an initial learning rate of 0.001 and a weight decay of 5e-4. (see Supp. for details).
Forecasting visual style We focus on forecasting the future of fashion over a 1- 2 year time course. In this horizon, we expect consumer purchase behavior to be the foremost indicator of fashion trends. In longer horizons, e.g., 5-10 years, we expect more factors to play a role in shifting general tastes, from the social, political, or demographic changes to technological and scientific advances. Our proposed approach could potentially serve as a quantitative tool towards understanding trends in such broader contexts, but modeling those factors is currently out of the scope of our work.
Style dynamics Having established the ability to forecast visual fashions, we now turn to demonstrating some suggestive applications. Fashion is a very active domain with styles and designs going in and out of popularity at varying speeds and stages. The life cycle of fashion goes through four main stages : 1) introduction; 2) growth; 3) maturity; and finally 4) decline. Knowing which style is at which level of its lifespan is of extreme importance for the fashion industry. Understanding the style dynamics helps companies to adapt their strategies and respond in time to accommodate the customers’ needs. Our model offers the opportunity to inspect visual style trends and lifespans. In Fig. 5, we visualize the temporal trajectories computed by our model for 6 styles from Dresses. The trends reveal several categories of styles: 1) Out of fashion: styles that are losing popularity at a rapid rate (Fig. 5a); 2) Classic: styles that are relatively popular and show little variations through the years (Fig. 5b); 3) Trending: styles that are trending and gaining popularity at a high rate (Fig. 5c and d); 4) Unpopular: styles that are currently at a low popularity rate with no sign of improvement (Fig. 5e); 5) Re-emerging: styles that were popular in the past, declined, and then resurface again and start trending (Fig. 5f). Our model is in a unique position to offer this view point on fashion. For example, using item popularity and trajectories is not informative about the life cycle of the visual style. An item lifespan is influenced by many other factors such as pricing, marketing strategy, and advertising among many others. By learning the latent visual styles in fashion, our model is able to capture the collective styles shared by many articles and, hence, depicts a more realistic popularity trajectory that is less influenced by irregularities experienced by the individual items.