SOMATCH · Project

AI-Powered Fashion Trend Prediction Tool Using Image Analysis and Social Media Data

digitalTestedTRL 6

Imagine you run a clothing brand and need to guess what colors, styles, and designs people will want next season — except instead of guessing, a computer scans millions of images from social media and online shops to tell you exactly what's trending right now. SOMATCH built software that does this automatically: it reads pictures of what people are actually wearing and buying, spots patterns, and shows designers clear visual trend reports. Think of it like a weather forecast, but for fashion — so companies can design what customers actually want instead of hoping they got it right.

By the numbers

consortium partners across 4 countries

SME partners involved in development and testing

75%

industry ratio in the consortium

total project deliverables produced

demo deliverables (alpha prototype + beta version)

The business problem

What needed solving

Fashion and textile companies lose money every season by designing products that miss consumer demand. Trend prediction today relies on expensive consultancies, trade show attendance, and subjective judgment — a slow, costly process that leaves SMEs at a disadvantage against large brands with bigger research budgets.

The solution

What was built

The team built a fashion trend prediction software platform that uses image analysis and social media mining to identify consumer preferences in real time. Concrete deliverables include an alpha system prototype for testing and a beta version with core functionality for end users, including mobile and wearable device support.

Audience

Who needs this

SME clothing and textile manufacturers needing affordable trend intelligenceFashion e-commerce platforms wanting to optimize product selectionIndependent fashion designers looking for data-driven design directionRetail chains seeking to reduce unsold inventory through better trend forecastingTrend forecasting agencies wanting to automate their data collection

Business applications

Who can put this to work

Fashion & Textile Manufacturing

SME

Target: SME clothing brands and textile producers

If you are an SME clothing manufacturer struggling to predict which colors, fabrics, and styles will sell next season — this project developed an image analysis tool that scans real consumer behavior online and delivers visual trend forecasts. Instead of relying on expensive trend agencies or gut feeling, you get data-driven design direction. The system was built specifically for textile and clothing companies with 5 SME partners testing it.

Fashion Retail & E-commerce

any

Target: Online fashion retailers and marketplace operators

If you are a fashion e-commerce platform dealing with high return rates because products don't match what customers actually want — this project built a real-time trend visualization system that tracks consumer preferences from social networks and open data. It helps you stock what people are already searching for and wearing. The tool includes mobile device integration for on-the-go trend monitoring.

Fashion Design & Creative Services

SME

Target: Independent fashion designers and design studios

If you are a fashion designer spending weeks researching trends through magazines, trade shows, and Instagram scrolling — this project created software that automates trend collection and visualization through advanced image analysis. It processes large sets of unstructured visual data and delivers actionable trend reports. The system was designed with wearable device integration including Google Glass for hands-free trend capture.

Frequently asked

Quick answers

What would it cost to use this trend prediction tool?

Based on available project data, no pricing model or licensing fees are disclosed. The project was an Innovation Action with 8 consortium partners, suggesting the technology was developed for commercial deployment. Contact the coordinator for current licensing or subscription terms.

Can this scale to handle large product catalogs and global trend data?

The system was designed to mine and visualize large sets of unstructured data from social networks and open data sources. The beta version delivered most core functionality for end users. Scalability to enterprise-level catalogs would need to be verified with the development team.

Who owns the IP and can I license this technology?

The consortium of 8 partners across 4 countries jointly developed the technology. With 6 industry partners and 5 SMEs involved, IP arrangements likely favor commercial exploitation. The coordinator, Universitat Politecnica de Catalunya (Spain), can clarify licensing availability.

How does this integrate with existing design and PLM software?

Based on available project data, the system was built as a standalone tool with dedicated interfaces for desktop and mobile devices. Integration with existing Product Lifecycle Management systems is not explicitly mentioned in the deliverables. The beta version focused on user experience and core trend analysis features.

What data sources does the tool actually analyze?

The tool processes images and data from social networks, e-commerce platforms, and open data sources related to consumer fashion preferences. It combines advanced image analysis technology with social network analysis to identify trends in colors, styles, and designs.

Is the technology still available or has development stopped?

The project ran from January 2015 to June 2016 and is now closed. A beta version and system prototype were delivered as part of the 9 total deliverables. Whether the software is still maintained or available for licensing would need to be confirmed with the consortium partners.

Consortium

Who built it

The SOMATCH consortium was heavily industry-driven with 6 out of 8 partners from industry and 5 being SMEs — giving it a 75% industry ratio. This is unusually high for an EU project and signals that the technology was shaped by real market needs rather than pure academic interest. The 2 university partners (TUM in Germany and UPC in Spain) provided the image analysis and content analysis expertise, while industry partners covered software development (Holonix, Sparsity, Ideal), end-user textile manufacturing and retail (DENA), and social networking and e-commerce (Weblogs, NJAL). The 4-country spread across Germany, Spain, Italy, and the UK gave the project access to major European fashion markets.

UNIVERSITAT POLITECNICA DE CATALUNYACoordinator · ES
HOLONIX SRLparticipant · IT
MANDELLI LAURAparticipant · IT
TECHNISCHE UNIVERSITAET MUENCHENparticipant · DE
I-DEAL S.R.L.C.R.participant · IT
SPARSITY SLparticipant · ES

How to reach the team

Universitat Politecnica de Catalunya (Barcelona, Spain) — search for the SOMATCH project lead in UPC's textile or computer science departments

Order a report on this consortium See UNIVERSITAT POLITECNICA DE CATALUNYA profile →

Next steps

Talk to the team behind this work.

Want to explore whether this fashion trend analysis technology could work for your business? SciTransfer can connect you with the research team and help evaluate fit.

Order a report on this topic More in Digital & ICT