Feature extraction and discrimination of fabric frictional sounds.

To discriminate different fabric frictional sound, a mechanism based on Haar features was established in this study, and an efficient approach was conducted to identify fabric friction sounds by discriminating against the performances of different fiber materials manifested in the mechanism. 4 categories of fabrics were put through a specialized apparatus to collect frictional sound signals. The Haar features on every scale and position of the acquired signal were extracted so as to establish a feature space. For each point in the feature space, a discriminator was built to approve all positive samples of a certain category and deny negative samples as many as possible. All negatives were filter out due to the probability of a single negative sample being allowed by a large number of discriminators is very small. The progressive selection was performed on the discriminators to form a queue so as to bring out a much reduced version of the unordered discriminators with the same discriminability. Results have indicated that the discrimination queue suggested by this study can substantially reduce the number of discriminators involved and meanwhile keep the discriminability of fabric friction sounds invariable. [ABSTRACT FROM AUTHOR]