26:T52f,

The granularity or speckle in medical ultrasound images tends to mask the presence of small lesions. As well, artifactual filling in of anechoic regions such as cysts, reduces the diagnostic potential of the images. These effects depend not only on the acoustic properties of the tissue but also are strongly influenced by the imaging system, especially the transducer geometry.

To study the effect of the transducer on the final B-scan image, a computer model has been developed simulating the interaction of ultrasound with a simple scattering medium. This model, incorporating the position dependence of the point response of the transducer, is based on single scattering from a collection of points positioned randomly in a three-dimensional volume. Using this approach, B-scan images showing speckle have been generated for different transducer geometries. Inclusion of a 2.6 mm void mimicking a cyst within the three-dimensional scattering volume has allowed us to predict the cyst contrast in the image for the different transducer systems.

Experimental B-scan images of a scattering phantom were obtained using the different pulse echo systems. Quantitative assessment using first and second order statistics of the images shows good agreement between experiment and theory.

27:T424,

The accuracy of the spectral difference method for measuring ultrasonic attenuation has been investigated using tissue-mimicking phantoms. Attenuation coefficients of the phantom materials were measured using a narrow-band substitution technique and compared with the results of the spectral difference method. Agreement within ±10 percent was typical for measurements in homogeneous materials. The best agreement between the spectral difference and substitution techniques was obtained when effects due to transducer beam diffraction were taken into account in the analysis. This was found for two types of homogeneous tissue-mimicking materials, both having speed of sound and attenuation properties similar to human liver but each with different backscatter properties. The effects of inhomogeneous tissues interposed between the transducer and the interrogated volume were also studied by simulating these conditions in phantoms. Experimental techniques which minimize the effects of perturbations introduced by these inhomogeneities are suggested.

28:T452,

A new method for solving the inverse scattering problem for the scalar, inhomogeneous, exact, Helmholtz wave equation is presented. No perturbation approximations are used and the method is applicable even for many cases where weak to moderate attenuation and moderate to strong refraction of incident fields occur. The ill-posed nature of the inverse scattering problem for a single monochromatic source is known. However, the use of multiple sources, the collection of redundant (i.e., overdetermined) data, and the constraining of the fields and complex refractive index to be spatially band limited constitutes a new problem. The cases we have tested by computer simulation indicate that the new problem is well posed, has a unique solution, and is stable with noisy data. The method is an application of the well-known method of moments with sinc basis and delta testing functions to discretize the problem. The inverse scattering solution may be obtained by solving the resulting set of simultaneous, quadratic, multivariate equations. Several algorithms for solving these equations are given.

29:T5b3,