Wavelet transform is able to characterize the fabric texture at multiscale and multiorientation, which provides a promising way to the classiÿcation of fabric defects. For the objective of minimum error rate in the defect classiÿcation, this paper compares six wavelet transform-based classiÿcation methods, using di erent discriminative training approaches to the design of the feature extractor and classiÿer. These six classiÿcation methods are: methods of using an Euclidean distance classiÿer and a neural network classiÿer trained by maximum likelihood method and backpropagation algorithm, respectively; methods of using an Euclidean distance classiÿer and a neural network classiÿer trained by minimum classiÿcation error method, respectively; method of using a linear transformation matrix-based feature extractor and an Euclidean distance classiÿer, designed by discriminative feature extraction (DFE) method; method of using an adaptive wavelet-based feature extractor and an Euclidean distance classiÿer, designed by the DFE method. These six approaches have been evaluated on the classiÿcation of 466 defect samples containing eight classes of fabric defects, and 434 nondefect samples. The DFE training approach using adaptive wavelet has been shown to outperform the other approaches, where 95.8% classiÿcation accuracy was achieved.