Brain stroke detection using continuous wavelets transform matching filters

Abstract

A computer-aided detection algorithm that is useful for diagnosing brain stroke is presented. Continuous Wavelet Transform matching filters is applied to identify the backscattered microwave signals from abnormal objects in the brain. The simulation results show that energy distribution of filters outputs in wavelets domain., in terms of time translation of correlation peak and range of scales in the estimated energy., is a promising candidate signature for diagnosing a hemorrhagic brain stroke. The simulation is done using an emulated head phantom., and two-dimensional finite-difference time-domain (FDTD).