The problem addressed by Somatch is a major issue that remains unresolved in the industry of high end clothing: consumer demands are unknown at the time that the clothing production is planned. The fact that the clothing manufacturers do not have a consistent and scientific insight but a rough and intuitive one (if present), on the users preferences on textiles, colours and shapes of clothes, results in a non-optimized design and production of new collections and in a very high rate of un-successful new design proposal. Design creation in fashion is very often lived as a pure creative process, where stylists act as pure artists, who interpret, on the basis of their personal intuitions, the trends of the market without any methodological support and with a high margin of risk. Only famous stylist, well established and working for large fashion groups, have the credibility and the advertising support to set their own trends into the market, while SMEs (90% of the total on the basis of Euratex estimation in 2011) have no access to their services and can have expectation of success only by a better interpretation of the market. If such information was available, clothing SME manufacturers would be able to drive their design process and to better tune production and retail choices in relation with trends and needs of consumers.
Therefore, the objective of the SOMATCH project was to improve the competitiveness of SMEs companies in the Textile and Clothing (T&C) sector by providing creative designers with detailed and reliable trends estimation and forecast of user acceptance. This way to achieve this goal was through the creation of an innovative tool for the mining and visualization of large sets of unstructured data related to the use and preferences of fashion products by consumers, supporting T&C companies quick reaction to the market dynamics and better adaptation of design to real consumers’ demand.”
Therefore, the main innovation objective of Somatch was the deployment of a tool for the generation and visualisation of knowledge from the huge, unstructured data related to textile and clothing (T&C), present in all suitable internet content, starting from image analysis.
SOMATCH faces this complex and challenging deal by the combined development and application of state-of-the-art advanced image analysis technology in clothing and fashion, as well as social network analysis.
The visualization of the generated data is performed from off-line statistics, generated after daily data processing, and by new real-time instruments for image collection and evaluation of designs. When commercialized, Somatch could open a vast field of new approaches for the fashion designers, supporting final user’s involvement into the whole trend evaluation and a close interaction with them.
To reach this purpose Somatch consortium included research centres expert in image and content analysis (TUM, UPC), software providers, experts in data management, platform development and fashion tools (Holonix. Sparsity, I-Deal), end users from SME textile industry and retail (DENA) and social networking, design community and e-commerce (Weblogs, Not Just a Label)
