A new method for grating and side lobes suppression in ultrasound images is presented. It is based on an analysis of the phase diversity at the aperture data. Two coherence factors, namely the phase coherence factor (PCF) and the sign coherence factor (SCF), are proposed to weight the coherent sum output. Different from other approaches, phase rather than amplitude information is used to perform the correction action.

Phase Coherence Imaging (PCI) has been recently proposed as a robust method to improve the quality of ultrasound images. Based on a statistical analysis of the instantaneous phase of the aperture data, a Phase Coherence Factor (PCF) is computed for every sampling instant. When used to weight the beamformer output, side and grating lobe levels are reduced and lateral resolution is increased. In this work, its application for grating lobe artifacts reduction is further investigated. Dependence of grating lobes reduction level with signal bandwidth and the number of array elements is analyzed and detection of low amplitude echoes located into the grating lobe region is addressed. Experimental data obtained with a 2-D matrix sparse array are presented. In a second experiment, a standard tissue-mimic phantom and a sparse linear array are used to evaluate the PCI grating lobe reduction performance for medical images.

Phase Coherence Imaging was proposed in 2009 as a new adaptive beamforming method, based on measuring the instantaneous phase dispersion of the aperture data. Low phase dispersion is indicative of a focused reflector, while large phase dispersion is produced by out of focus reflectors, reverberations and the grained structure of some materials. Based on a phase dispersion metric, a coherence factor can be defined and used to weight the beamformer output, in such a way that focused reflectors are maintained (low phase dispersion), while side-lobes, grating-lobes, reverberations and grain noise are reduced (large phase dispersion). Besides increasing the image dynamic range by reducing the sidelobes and background noise, lateral resolution is simultaneously improved by reducing the main-lobe width.Several phase coherence factors have been defined, based on different dispersion metrics with different computational complexity and noise reduction levels. In particular, a realization based on the sign of the received signals (Sign Coherence Factor) allows implementing the method in real-time with very low hardware resources requirements. The coherence factor can be used to weight the beamformer output or as an image itself, with the property of representing reflectors with amplitudes independent of their reflectivity.

This work analyses the use of parallel processing techniques in synthetic aperture ultrasonic imaging applications. In particular, the Total Focussing Method, which is a O(N ² x P) problem, is studied. This work presents different parallelization strategies for multicore CPU and GPU architectures. The parallelization processes on both platforms are discussed and optimized in order to achieve real-time performance.