In recent years, artificial intelligence (AI) has been increasingly adopted across the textile value chain. Within textile characterisation, AI is now being used to augment both objective instrumentation (e.g. mechanical, spectroscopic, imaging, and thermal methods) and subjective assessments of fabric handle and comfort. This review focuses on (i) characterisation modalities commonly used for fibres, yarns, fabrics (woven/knit/nonwoven), and post-consumer garments; (ii) AI task families relevant to characterisation; and (iii) how AI can support interpretation, automation, and decision-making for FAST/KES-type mechanical tests, FTIR, SEM, and thermal analysis (DSC/TGA/DMA). AI for fashion design, trend forecasting, and purely creative generative applications is outside the scope unless directly linked to characterisation outcomes. Beyond a broad overview, the review contributes a unified taxonomy that maps characterisation modalities to AI tasks and data types, synthesises cross-cutting limitations, and proposes practical reporting and benchmarking recommendations to improve reproducibility and comparability. Finally, implications for textile recycling are discussed, where robust material identification and contamination detection can enable higher-quality sorting and recovery in circular systems.
Azeem et al. (Fri,) studied this question.