Adaptive beamforming has been widely used as a way to correct phase and amplitude aberration errors in medical ultrasound. A less-studied concern in ultrasound beamforming is the deleterious contribution of off-axis bright targets. We describe a new approach, the constrained adaptive beamformer (CAB), which builds on classic array processing methods. Given a desired frequency response in the focal direction, the CAB dynamically imposes an optimal set of time-dependent weights on the receive aperture, reducing signals from directions other than the focal direction. Two implementations of the CAB are presented which differ in their use of calculated weights to form an output image: the Single Iteration CAB and Multiple Iteration CAB.
We present results from experiments performed on a Philips SONOS 5500 imaging system operating with an 8.7 MHz linear array and contrast the performance of the two CAB implementations. Data was acquired from wire targets in a water tank and low echogenicity cysts in a grayscale tissue mimicking phantom. The desired system frequency response was specified by a FIR filter with the same center frequency as the transducer. Improvements in lateral resolution for wire targets and contrast for low echogenicity cysts are shown. Simulations are used to demonstrate limitations of the CAB.