Advantage of automatic fabric inspection include consistent, objective, and reproducible assessment of fabrics; high testing speed; direct electronic data accumulation for quicker transmission and processing; and fault classification that allows assessment of the fabric faults irrespective of the end product. The use of neural networks in the Uster Fabriscan fabric inspection system allows it to simultaneously learn the characteristic appearance of flawless fabric and expected characteristics. The system can also memorize deviations that are not to be characterized as faults. The system allows fabric manufacturers to identify quickly and adjust any loom that produces faults. 4 refs.