Time frequency analysis of temporomandibular joint sounds using Gabor expansions

In this paper we present an evolutionary spectral method for the analysis of temporomandibular joint (TMJ) sound signals. The analysis of TMJ sounds is of great interest in oral and dental research areas as an alternative technique for the diagnosis of TMJ disorders to the conventional methods. Here, an evolutionary spectral estimate of a TMJ signal is obtained using the multi-window Gabor expansions that use scaled windows modulated by linear chirps as basis functions. We show that this method performs better than the short-time Fourier transform and the reduced interference distribution in representing the TMJ signals.