Enhancing ultrasound tissue characterization with the double Nakagami distribution model

The single Nakagami distribution (SND) models backscatter envelope statistics for tissues with a single type of scatterer but is less effective for tissues containing multiple scatterer types. This study introduces a new Double Nakagami distribution (DND) estimator to model tissues with two types of scatterers. The DND model directly estimates three parameters corresponding to number density and relative volume ratio of the first scatterer type, from which two additional parameters corresponding to the second scatterer type are derived through a method-of-moments approach. An algorithm was developed for efficient model initialization. To validate the method, simulations were conducted by generating envelope statistics data with various combinations of the five DND parameters. Experimental data were also collected from agar phantoms containing varying concentrations of nylon and acrylic spheres. The simulations showed the DND model parameter estimation was more accurate than the SND model with errors below 5%. For the phantom tests, the DND model accurately estimated scatterer proportions with errors below 6%. The DND fit model demonstrated significant advantages in terms of computational accuracy for tissues containing more than one type of scatterer. The DND model has the potential to improve quantitative tissue characterization when there are multiple scatterer types.