World-unique AI database revolutionises second-hand sorting, here’s how

Wargön Innovation is an innovation hub dedicated to creating a circular and sustainable society. It’s about providing support to entrepreneurs, businesses, and public partners striving to develop and promote groundbreaking products, processes, and business models. The hub just secured additional funding from the European Regional Development Fund for a three-year initiative aimed at assisting small and medium-sized businesses in adopting circular business models. 

— As a result, we will be further scaling up our operations throughout the autumn to extend our support to a broader range of participants, says Susanne Eriksson, project manager. She continues:

— By collaborating with us, innovators, business leaders, and academic experts have the opportunity to explore new technologies and methods. These advancements pave the way for circular business models and contribute to a resource-efficient textile and fashion sector.

The hub’s core competencies lie in orchestrating circular processes, leveraging textiles and bio-based materials, and pioneering industrial textile sorting.

— Notably, we’ve established a unique testing and demonstration facility in these areas, enhancing accessibility and driving progress.

Susanne Eriksson has experience working with sustainable materials and adherent process and production technology. She’s also responsible for the project AI for resource-efficient circular fashion, a collaboration with RISE Research Institute of Sweden, ShareTex, and second-hand industry players in Sweden and Switzerland.

— The main purpose is to identify, design, and develop data-driven AI for textile second-hand, with a special focus on automating the sorting process for textile reuse, she explains. The goal is to create an open database of used clothes and then use this for machine learning that can facilitate the work of textile sorters. The overarching objective is to significantly extend the user phase of clothing so that new production is replaced.

— As the initiator and project manager, we have contributed expertise in sorting and valuation of second-hand items, as well as knowledge for designing and testing the AI tool, and work on life cycle analysis. We’ve also been responsible for designing and installing specially designed photo stations.

The second-hand market is growing rapidly in Europe, but the capacity to collect and sort clothes is severely undersized. At the same time, the European Commission has taken further steps towards a more sustainable and circular textile industry.

— This includes the proposed law suggesting new rules that make producers responsible for the entire life cycle of textile products and therefore obligated to finance collection and sorting that enables reuse and recycling. Part of the solution is a world-unique open AI database for intelligent sorting and valuation of second-hand items developed within our project.

— The sorting capacity in Europe is only sufficient to handle a quarter of what is collected today. The rest is incinerated or ends up in landfills, even though as much as 50 to 75 per cent of the collected textiles are possible to reuse. With the help of smart technology and automation, an efficient and high-tech multi-billion-dollar industry is opening up within the second-hand market.

— For clothes to be sold second-hand, manual sorting and valuation are required, which is physically demanding and expensive work. Additionally, it is difficult for a human to assess each garment in a short amount of time. With the help of artificial intelligence and machine learning, tens of thousands of different garments can be recognised, and those evaluating the items receive a solid basis for deciding whether the items should go to second-hand or not. The use of AI in second-hand sorting significantly relieves the sorting personnel. Both the speed and accuracy of automation have the potential to far exceed what human handling can achieve. The hope is that the database can be used by the many actors who, in various ways, sell or rent out second-hand clothes today.

The database created within the project, Eriksson explain, is world-unique as it is the first completely open database for second-hand items. 

— Today, there are many open databases for newly produced clothes; for example, AI is used to enhance online sales of newly produced garments. Nothing equivalent exists for second-hand clothes — until now. An open database means that all data is available for anyone wishing to build their own algorithms. By creating an open database, the project aims to support the digital development in textile sorting and thereby encourage the use of the data and the development of new digital solutions.

— Within the project, photos have been taken in the mentioned, specially designed photo stations located at Wargön Innovation, as well as at second-hand stores’ sorting facilities. The photo stations have scanned second-hand garments to create a completely unique open database. In concrete terms, this means that garments have been photographed, and information about patterns, materials, brand, level of wear and tear, possible price, and fashion degree has been saved in the database.

What’s been the hardest part when developing it?

— This project is a first step in testing how the database and AI tool function. Several areas for further development and opportunities for future projects have already been identified. Textile sorting is moving towards large-scale operations, and those require automation. The textile material’s many different variables in properties make AI a very useful tool that can be adapted to the situation and needs. However, the database will not only be useful for large-scale operations but also for smaller actors who can build algorithms based on their specific needs.

During this autumn, tests of the tool will be conducted by the stakeholders — the second-hand actors — so that they can provide feedback on development possibilities.

— The database will also continue to be built continuously to reach 30,000 items. AI can be used to recognize many different properties, but it requires data, and building up data for all the properties one wants to use AI for is a challenge. A challenge that we are happy to take on!

What else is on your mind now?

— Given the more defined legal requirements associated with textiles, coupled with the swift evolution of climate and environmental changes, it is important for the industry to collaborate swiftly to attain sustainable and circular practices, Eriksson shares. She continues:

— A critical emphasis lies in the establishment of sustainable production practices that leverage existing materials. To achieve this, it is significant to establish processes for sorting and evaluating textiles that are no longer in use.

— A robust and well-functioning sorting infrastructure is necessary to handle the substantial volume of textile materials in circulation and enable their continued use through methods like reuse, redesign, repair, and recycling. Accelerating the progress of textile sorting is crucial to prevent the sorting stage from becoming an obstacle to the industry’s transition.