The world is entering its 4th Industrial Revolution, a new era of manufacturing characterized by ubiquitous digitization and computing. This is true also for the fashion industry, an industry in which Europe (and Italy in particular) has long maintained a global lead. IT-SHIRT is driven by the argument that, for the European fashion industry to retain its global lead, it needs to make digital technologies an integral part of the fashion value chain. In this research, we describe concrete ways in which such an integration can be achieved. Compared to the traditional approach that is based entirely on physical models, the digital aspect becomes an accelerating factor that facilitates collaboration between designers, minimizes the time-to-market, and enables immediate consumer feedback. Specifically, our project aims at investigating (i) CAD tools for real-time collaboration of designers over virtual 3D models of
fashion garments, (ii) software engines on top of CAD tools to provide designers with smart recommendations and creative support, (iii) online and offline digital showroom platforms, and (iv) social marketing engines for capturing consumer feedback and providing targeted marketing recommendations. The aforementioned core-components weave a "digital thread" around the production pipeline of the fashion industry, that spans the stages of conception, production, distribution, access to audiences, and market feedback. To assess the practical utility of the solutions investigated during the project, we will validate it on the challenging Italian fashion market and in international fashion initiatives through ALTAROMA, in collaboration with fashion designers and consumers.
From the scientific point of view, IT-SHIRT addresses several research areas which will be presented by theme, following the technical objectives previously identified.
Objectives 1 & 3: IT-SHIRT Fashion CAD solution and Fashion Digital Showroom Shops
3D collaborative modeling. IT-SHIRT goals for collaborative garment design are (1) that designers edits are faithfully maintained, (2) that collaboration is interactive, and (3) that concurrent edits are merged robustly to ensure that the garment remain consistent, and (4) that the system is reliable against network failures and users participation in the collaboration. To achieve all these goals, we cast collaborative garment editing as an instance of distributed version control.
Interactive animation of virtual garment. IT-SHIRT will identify, develop and validate new realtime animation technology for 3D virtual clothing usable by industry. The obtained software shall solve the dynamics of complex textile garments in less than 2ms per animation frame. This is the limit required to employ devices for Virtual Reality like the Oculus Rift, which is likely to be used by the industry design and entertainment in the near future.
Multidevice interaction for 3D. IT-SHIRT will combine the design outcomes and empirical findings from the preceding parts to compile and validate guidelines for fixed and mobile collaborative surfaces. Our work is motivated by the roles smartphones and tablets play in everyday interactive spaces as well as anticipated
developments in mobile sensing and haptic output technology (Alvina et al., 2015).
Objective 2: Enhancing creativity of designers at concept and modeling level
Mixed Initiative Design. IT-SHIRT will extend current state-of-the-art CAD tools for fashion design and enhance them with more proactive computational input. The computational input takes the form of suggestions and alternatives, shown to the designers during their design process. As an extension to the current literature, the mixed initiative design tools of IT-SHIRT will identify the use cases of CAD tools for fashion designers and explore how suggestions can be presented in a way that fosters rather than blocks human creativity in the design process. Moreover, IT-SHIRT will explore the algorithms used by the computational initiative, which range from presenting appropriate, themed existing fashion styles to algorithmically generated, out-of-the-box designs.
Personalized, gameful interactions. Accounting for the criticisms of scholars against meaningless, extrinsic motivation via point systems, IT-SHIRT will expand on existing frameworks of gamification by exploring methods of increasing intrinsic motivation - this can be achieved by personalizing and matching the difficulty or goals of the playful interface and its users. Moreover, it will explore the construction of stylistically coherent groups of users in order to enhance the collaborative creativity of its members. As a further step for personalizing the user's interaction with these gameful environments, IT-SHIRT will expand on the current state-of-the-art on preference modeling in order to adapt it to the needs of fashion specific aesthetics. Such aesthetics will be derived from fashion design principles (theory driven) as well as from data collected from publicly available data (e.g. social media). The learned preferences using these algorithms will be used to personalize the interaction between users and the gameful environment. Moreover, the learned preference will inform both recommender systems of existing designs as well as act as criteria for providing new designs tailored to the user's tastes.
Objective 4: Capturing and evaluating consumer feedback and allow for targeted social marketing
Data driven analytics for detecting Fashion Trends. Trend detection in the vertical domain of fashion poses challenges that have not been addressed by the techniques developed so far. One challenge is that fashion related activity on social media usually involves images and therefore it falls outside the scope of already developed trend detection techniques that have focused on the easier case of text analysis. Moreover, fashion related images posted on social media are typically accompanied with url links that point back to the source of those images, typically a website belonging to a fashion magazine, retailer, or designer. Such websites contain information about the fashion items pictured in the posted image and their analysis can provide valuable insights for the purposes of trend detection.