Computer phantoms for simulating ultrasound B-mode and CFM images

Programs capable of simulating ultrasound images have recently been developed. This opens the possibility of evaluating transducers and focusing schemes not only from their point spread function, but also from an imaging point of view. The calculation of the ultrasound field is based on linear acoustics using the Tupholme-Stepanishen method for calculating the spatial impulse response. Any transducer can be simulated by splitting the aperture into rectangular or triangular sub-apertures, and the calculation can include any transducer excitation and apodization. The acoustic settings can be controlled in the entire image through dynamic apodization and focusing. The transmit and receive apertures can be defined independently of each other. Frequency dependent attenuation can also be included in the simulation. The B-mode images are generated by specifying a number of independent scatterers in a file that defines their position and amplitude. Adjusting the number of scatterers and their relative amplitude yields the proper image. Five different computer phantoms are described. The first one consists of a number of point targets. It is used for studying the point spread function as a function of spatial position, and can give an indication of sidelobe levels and focusing abilities. The second phantom contains a number of cysts and point tagets along with a homogeneous speckle pattern. This is used for investigating image contrast, and the system's ability to detect low-contrast objects. The third phantom is for realistic clinical imaging. It contains the image of a 12 week old fetus, where the placenta and the upper body of the fetus is visible. This phantom gives an indication of the whole system's capability for real imaging. The current fetus phantom is only two-dimensional, as it is constant in reflection amplitude in the elevation direction. The program, however, can handle the full three-dimensional simulation, and the whole body could in principle be simulated. An example for a simulated kidney is also shown. The last phantom is used for color flow mapping and is a combination of static and moving scatterers. A model with stepwise movement of the scatterers in an artery with a parabolic how profile surrounded by tissue is used. The signal from the scatterers is recorded between each movement and then they are propagated for the next image acquisition. The phantom can be used to study both spectral estimation techniques and color flow mapping estimators. The scatterer description is three-dimensional and so is the acoustic field calculation. This approach allows the evaluation of an imaging system design to be performed very fast and early in the development process. Typically a single phantom simulation takes less than 12 hours of simulation time. Transducer designs can be optimized and the practical implementation reduced to only one trial. Also different signal processing approaches can be evaluated realistically.