By making textile-fibre identification faster, open and machine-readable, the FasTEX project attacks a quiet bottleneck in recycling, pollution tracking and forensic science.
Researchers at King’s College London have published an open reference library of infrared spectral “fingerprints” for textile fibres, designed to improve how fibres are identified using FTIR spectroscopy and machine learning. The associated Data in Brief paper describes a dataset built from 137 textile samples covering 26 natural and man-made fibre types, together with metadata, processed reference data and reusable analysis code.
The project, called FasTEX (“FAST identification of TEXtile fibres”), addresses a persistent problem: fibre identification is often slow, expert-dependent and hard to standardise across labs. King’s says the new dataset was created under standardised laboratory conditions so other researchers can benchmark classification methods on a common base rather than build isolated in-house libraries.
This is not just a laboratory convenience. Reliable fibre identification is increasingly central to microfibre pollution research, forensic investigations, and textile recycling, where sorting accuracy determines whether material stays in a higher-value loop or drops into waste. Existing libraries are often incomplete or inaccessible; openness is therefore part of the innovation.
The real value lies in scale. If openly shared spectral libraries become robust enough, fibre identification could move from specialist interpretation toward reproducible, automated classification. That would make circular-textile systems less dependent on human bottlenecks—and more industrially credible.
