•New

Eight Catalan, Spanish, Italian, English and German companies and technology centres, coordinated by the UPC's Innotex Centre/INTEXTER, have begun to develop an innovative smart system called SOMATCH that is able to predict the fashion trends that are likely to be most successful by processing billions of images that have been tracked and viewed on social networks according to specific parameters. SOMATCH will principally analyse images published on Twitter, Instagram and the blogs of some of the world's most influential figures in fashion.

Once it has been developed, SOMATCH will enable small and medium-sized enterprises (SMEs) that are engaged in textile design to access the restricted field of objective predictions of fashion trends, a field to which only the industry's major brands, with large human resources, currently have access. The project's managers anticipate that using the SOMATCH system could save European SMEs €8,000 million a year.

The large fashion companies have armies of trendhunters who travel around the world analysing colours, fabrics and designs of the clothes we wear. This allows them to create fashion trends and confidently produce their design lines every season.

However, 90% of the fashion industry in Europe consists of SMEs, which have to take risks on their clothing designs up to 18 months before they are put on sale. Therefore, predicting the colours, designs and fabrics that will be worn each season is essential--if they don't get it right, they could lose a lot of money. In fact, SMEs in the clothing sector in Europe (numbering about 146,000 in 2012 according to Eurostat) have to invest between 6% and 7% of their turnover in production design, totalling about €11.5 billion.

Aware of this fact, researchers at the UPC's Innotex Centre/INTEXTER submitted to the European Union, with the support of seven Italian, German, English and Spanish companies and technology centres, a project to create a smart system that would allow small and medium-sized European clothing companies to objectively predict fashion trends with enough time to support their design decisions.

The project is called SOMATCH and it consists of a robot or software that is capable of analysing billions of images from social networks to obtain objective indicators for predicting future fashion trends. In compliance with the European data protection law, one of the EU's conditions for allowing the project to go ahead was that the images captured by the system will not be stored.

The data analysed by SOMATCH will distinguish between men's and women's clothing and the approximate age of the wearers. The type of garment (T-shirt, shirt, skirt or trousers) and style (tight, short, long, colours, combinations and textures) will also be identified.

The billions of images that are available can be used to generate very refined and reliable statistics sorted according to the needs of the designers. Therefore, companies in the sector can plan their production on a sounder basis.

José Antonio Tornero, coordinator of the SOMATCH project, says "to make a shirt, a small business may perform up to 66,000 virtual combinations of yarn, textures and colours. "Of these 66,000 possible combinations, only 3% are selected for collections. Of this 3%, only half, i.e. 1,000 combinations, are finally selected for manufacture, and only 500 different types of shirt end up in the shops", says Tornero. So only 500 are left of the initial 66,000 designs. "This process involves an extraordinary effort for the 60% of companies in Europe that have fewer than 50 employees", says the researcher.

The SOMATCH project could save European SMEs in the textile sector up to €8 billion by drastically reducing and simplifying the design production process, in which these companies currently invest nearly 7% of their turnover.

SOMATCH will be ready in 18 months and has a budget of €1.2 million. The project, developed under the Horizon 2020 programme, is coordinated by the UPC's Innotex Centre/INTEXTER. Sparsity, a UPC spin-off that is involved, is responsible for analysing social networks to detect trends using a system for managing a high-performance graph database. Also involved are the companies Holonix, i-Deal and Dena Cashmere from Italy, Not Just a Label of the UK, Weblogs SL of Madrid and the Technische Universität München.

According to Eurostat 2012, of the 146,000 European textile manufacturing companies, 90% are SMEs employing nearly two million people. Of these companies, 60% have fewer than 50 employees.

Small and medium-sized textile manufacturing enterprises have a domestic turnover of €178 billion in Europe, where domestic consumption is more than €500 billion. The sector generates €38 billion in exports.

The Innotex Centre is a powerful R&D centre consisting of three major university research groups in the field of textiles: the Institute of Textile Research and Industrial Cooperation (INTEXTER), the Technological Innovation Centre (CTF) and the Toxicology Research and Innovation Centre (CRIT), all three of which are based on the Terrassa Campus.

The technological capacity of the Innotex Centre allows it to serve national and international businesses in the textile sector and other industrial sectors. It is a member of the TECNIO network of the Government of Catalonia and is part of the UPC's Innovation and Technology Centre (CIT UPC), the University's technology centre.