The second-hand market is experiencing rapid growth in Europe, yet the capacity for collecting and sorting clothing falls significantly short of the demand. Concurrently, the European Commission is taking strides towards a more sustainable and circular textile industry, proposing new regulations that make producers accountable for the entire lifecycle of textile products. This obliges them to finance the collection and sorting processes that enable reuse and recycling. A critical component of this solution is a groundbreaking open AI database, created for the intelligent sorting and valuation of second-hand items, as part of the AI for Resource Efficient Circular Fashion project. This initiative is a joint effort between Wargön Innovation and Research Institutes of Sweden (RISE)
Today, Europe’s sorting capacity is only sufficient to handle a quarter of the items collected. The remainder is either incinerated or sent to landfills, despite the fact that as much as 50 to 75 percent of collected textiles can be reused. Leveraging smart technology and automation presents an opportunity for a highly efficient, high-tech second-hand industry, worth billions of Swedish crowns.
While there are numerous open databases for newly produced clothing that utilize Artificial Intelligence (AI) to enhance online sales, no equivalent exists for second-hand clothing—until now.
“This database is truly revolutionary! It’s the first of its kind capable of analyzing multiple attributes of second-hand clothing, from material composition to identifying stains and tears. It’s not just data; it’s a catalyst for sustainable fashion!” remarked Farrukh Nauman, an AI researcher at RISE.
Assessing and sorting second-hand clothing has traditionally been a labor-intensive and costly endeavor, and humans struggle to evaluate each garment quickly. With the assistance of AI and Machine Learning (ML), tens of thousands of diverse garments can be identified. Those evaluating the garments now have a strong foundation for determining whether they should be resold.
“The application of AI in second-hand clothing sorting offers substantial relief for sorting personnel. Both the speed and accuracy of automation have the potential to surpass human capabilities. We hope this database can be utilized by numerous stakeholders who are involved in selling or renting second-hand clothing in various ways,” says Susanne Eriksson, project manager at Wargön Innovation and head of the AI for resource-efficient circular fashion project. “By establishing an open database, we aim to support the digital evolution in textile sorting and encourage the use of data to create innovative digital solutions,” she added.
In the AI for resource-efficient circular fashion project, thousands of images have been captured in specially designed photo stations at locations like Wargön Innovation, Björkåfrihet, and Myrorna’s sorting facilities. These stations have been scanning second-hand garments to create a unique open database. In practical terms, this means that garments have been photographed, and information about patterns, materials, brands, texture, possible pricing, and fashion relevance have been recorded in the database.
“With AI, second-hand retailers can generate comprehensive descriptions of garments and receive recommendations on the best platform or store for selling each unique item,” explains Farrukh Nauman.
“This project is just the beginning, a testing ground for the database and AI tool. We’ve already identified various areas for further development and opportunities for future projects. The scale of textile sorting is expanding, and large-scale sorting demands automation. The multitude of variables in textile material properties makes AI an excellent tool adaptable to specific situations and requirements. However, the database will not only benefit large-scale operators but also smaller entities that can create algorithms tailored to their specific needs,” adds Susanne Eriksson.
One of the project’s participating companies is the second-hand store Inimini. Josefine Runquist, the founder of Inimini, is enthusiastic about the open database:
“It’s exciting to witness the convergence of second-hand and AI. This development bodes well for the future. If AI can make second-hand commerce profitable, more companies will undoubtedly embrace this sustainable trend.
ShareTex and its CEO, Miguel Sanchis Sebastia, also see the positive potential of AI in second-hand sorting and valuation:
“This project provides an ideal opportunity to strengthen our collaboration with Wargön Innovation and align sorting technology development with the needs of textile recyclers. The AI database can serve as a quality assurance for our raw materials, ensuring that we’re not recycling materials that can be repaired or resold.”
Further Development and Refinement
The project, scheduled to run until April 2024, will continue to photograph and analyze garments to expand the database with more detailed information. The AI algorithms and user interface will also undergo further refinement to simplify the process of sorting, selling, and shopping for second-hand items.
“Throughout the autumn, the tool will be tested by the end-users (second-hand actors) to gather feedback and identify development opportunities. AI can recognize various characteristics, but amassing data for all the characteristics intended for AI application presents a challenge. It’s a challenge we’re eager to embrace!” concludes Susanne Eriksson, project manager at Wargön Innovation.
The AI database is developed within the AI for Resource-Efficient Circular Fashion project, a collaborative effort involving Wargön Innovation, RISE Research Institute of Sweden, ShareTex, Inimini, and various players in the second-hand industry such as Myrorna, Röda Korset, Björkå Frihet, and Texaid (Switzerland). The project is funded by Vinnova, and the database is being further adapted for large-scale sorting as part of the CISUTAC (Circular & Sustainable Textiles & Clothing) project.