Introduction

Creative domains render a big part of modern society, having a strong influence on economy and cultural life. Much effort within creative domains, such as fashion, art and design, centers around the creation, consumption and analytics of creative visual content. In recent years, there has been an explosion of research in applying machine learning and computer vision algorithms to various aspects of the creative domains including generating, analyzing and processing visual content. This ever-increasing interest is most evident in two important research trends: (1) Computer Vision for Fashion and (2) Visual Content Generation for Creative Applications.

This workshop aims to bring together researchers in computer vision and machine learning and artists and professionals from creative domains to discuss open problems in the areas of computer vision for fashion and creative visual content generation. We provide diverse workshop programs to encourage interdisciplinary discussions and idea sharing, including six keynote talks, a paper submission track, a juried art gallery, and a challenge on fashion image generation.

Announcement

• Accepted papers are listed here.
• Accepted artworks are available here.
• The social mixer of the workshop is on Sep. 12th
• FashionGen challenge deadline: July 31st
• Notification of paper acceptance: July 30th
• Notification of art competition acceptance: July 19th

Participation

Paper Submission

We solicit paper submissions on novel methods and application scenarios of CV for creative applications with a focus on fashion and visual content generation. We accept papers on a variety of topics, including generative models, retrieval, product recommendation, image segmentation, attribute discovery and trend forecast, etc. Papers will be peer reviewed under double-blind policy and the submission deadline is 15th July 2018. Accepted papers will be presented at the poster session and one paper will be awarded as the best paper. Please see more information on paper submission here.

Call for Artworks

We will have an artistic jury consisting of a committee of artists, curators and technologists working at the intersection of computer vision, ML and art. They will both accept and reject submitted artworks on the basis of aesthetic value, artistic contribution and technical interest. We will show all the accepted works in the gallery show, and further award a best in show prize to the most outstanding contribution. Artwork documentation and statement submission deadline is 19th July 2018.

Challenge: Fashion Image Generation

We introduce the FashionGen dataset, which is a novel dataset of hundreds of thousands of high quality fashion images. The dataset consists of multiple angles of each product against a standardized background, worn by a model under consistent lighting conditions, along with metadata such as text descriptions of the products. The challenge is to generate images conditioned on the text descriptions. The quality of the generated images will be evaluated both by a pre-trained network that will be exposed via an API to the contest participants and by human evaluators. We will accept submissions based on both performance on the quantitative benchmark, as well as the qualitative nature of the results. The best submission will receive an award and will present their work at the workshop. For more information on the challenge, please visit this page.

Invited Speakers

Sponsors

Contact

• For questions, please contact the workshop chairs, Hui Wu and Negar Rostamzadeh.