Wavelets and Zipf law for audio signal analysis

We present in this paper an audio signal classification method based on the selection of features to be chosen among a larger set. We have included in the possible features some that are deduced from the Zipf and inverse Zipf laws. These laws are powerful analysis tools allowing the extraction of information not available by standard methods. In order to apply Zipf and inverse Zipf laws on audio signals, we propose two ways of coding these signals, based on their temporal and time-scale representations. The most pertinent features linked to Zipf and inverse Zipf laws are then selected by a method based on a genetic algorithm. Finally, the classification step aims at the identification of signals. Four classification methods have been considered as well as a fusion method used to combine these classifiers. In order to evaluate our method, we have analysed medical signals corresponding to swallowing signals containing xiphoidal sounds. The problem is to characterize them according to the gastro-oesophageal reflux pathological state.