Adaptive imaging using principal-component-synthesized aperture data

Abstract

Sidelobe contribution from off-axis scatterers degrades image quality in ultrasound imaging. Focusing errors resulting from sound-velocity inhomogeneities in tissues, also known as phase aberrations, reduce coherence of the received signals and thus elevates the sidelobe level degrading the contrast resolution. In this study, we proposed a novel adaptive sidelobe-reduction technique using aperture data synthesized from mainlobe-dominant principal components. Principal component analysis of aperture data with proper delay being applied is performed here by singular value decomposition where the singular values are in order of large to small. These singular values describe how much energy of the aperture data is accounted for by the associated principal component. Because of the high coherence, the mainlobe-contributed energy in the aperture data concentrates on the principal components associated with the first few singular values whereas the one from sidelobe contribution spreads over all the principal components. In our method, the first few principal components are used to synthesize a new aperture data where undesired sidelobe contributions are partly reduced and then beamsum is performed on this new aperture data; thus reducing the effect of focusing errors and compensating the degraded image quality. Real array data are used to evaluate the efficacy of the proposed technique. The proposed technique offers contrast enhancement from 0.4 (no distortion) dB to 5 dB (2x distortion) and improvement of contrast-to-noise ratio from 2.5 % (no distortion) to 24 % (2x distortion) in real ultrasound data. It is demonstrated that the proposed technique effectively reduces sidelobes contribution and thus improves the contrast resolution.