Poster Submissions

 

Fashion and Apparel Browsing for Inspirational Content

Annette A. Ward, Stephen J. McKenna, Ian W. Ricketts, Ruixuan Wang, Junwei Han, Wei Jia, University of Dundee; Paul Sergeant, Calico Jack; Anna Buruma, Peter Taylor, Liberty Fabrics; Mike Stapleton, Mike Coyne, Michael Selway, Graham Howard, System Simulation Ltd.; James Stevenson, Victoria & Albert Museum; Chris Wilkie, Nic Sheen, Advisory Members.

Designers use visual images to spark and fuel their creativity. The Fashion and Apparel Browsing for Inspirational Content (FABRIC) project is developing and evaluating software that can help such creative processes by enabling effective browsing, retrieval, and management of digital art images throughout the textile and clothing industry, with application to other creative industries. This 3-year 1.4M Technology Strategy Board project, sponsored by the Department for Innovation, Universities and Skills (DIUS), and led by the University of Dundee, includes collaborators Liberty Fabrics, the Victoria & Albert Museum, Systems Simulation Ltd, and Calico Jack. Objectives include real-time browsing on desktop and mobile platforms, provision of a pre-market software prototype, in-depth performance evaluation, a browsing engine, and an intuitive user interface for visualisation and navigation of image manifolds. Computational methods for visualization of image collections based on manifold learning have been developed to make effective use of limited display size. Visual content description makes use of a novel Bayesian method for regular texture analysis and Markov random field-based segmentation. The software is being developed and evaluated using the collections at Liberty Fabrics and the Victoria and Albert Museum. Both are design icons; Liberty, as a design business and the V&A as a repository of design. Both sell their images and image rights to individuals and product developers. Visual access to those designs means easier access and potential for increasing revenue.



Hosted by the Knowledge Media Institute