FPGA realization of an automatic wheeze detector based on wavelet filter bank

This study presents the design, simulation and implementation of a portable wheeze detector based on a four-level Daubechies 4-tap orthogonal wavelet filter bank. The proposed wavelet filter bank is applied to decompose the lung sounds into different frequency bands. The relative wavelet energy of the fourth high-pass level wavelet decomposition is used as an indicator to identify wheezing episodes. The simulation results indicate that the relative wavelet energy is capable of providing feasible performance in recognizing wheeze sounds. Afterward the wavelet filter bank is realized using the lattice filter architecture. The architecture is implemented using field programmable gate array technology.