Non-Rayleigh statistics of ultrasonic backscattered echo from tissues

Abstract

The envelope of the backscattered signal from tissues can exhibit non-Rayleigh statistics if the number density of scatterers is small or if the variations in the scattering cross sections are random. The K distribution which has been used extensively in radar, is introduced to model this non-Rayleigh behavior. The generalized K distribution is extremely useful since it encompasses a wide range of distributions such as Rayleigh, Lognormal, and Rician. Computer simulations were conducted using a simple one-dimensional discrete scattering model to investigate the properties of the echo envelope. In addition to cases of low number densities, significant departures from Rayleigh statistics were seen as the scattering cross sections of the scatterers become random. The validity of this model was also tested using data from tissue mimicking phantoms. Results indicate that the density function of the envelope can be modeled by the K distribution and the parameters of the K distribution can provide information on the nature of the scattering region in terms of the number density of the scatterers as well as the scattering cross sections of the scatterers in the range cell. [Work was supported by NSF Grant No. BCS-9207385.]