Ole Marius Hoel Rindal will present his PhD work he has done on software beamforming in medical ultrasound imaging.
Abstract: Medical ultrasound image reconstruction in software, known as software beamforming, has revolutionized the flexibility of the methods used to create an ultrasound image. Software beamforming is possible due to the improvements in the processing power of available hardware. Software beamforming is a blessing for the image quality since it provides access to the raw data early in the processing chain. Early access to the data means that the image reconstruction can better exploit the data through, e.g., adaptive beamforming. Another blessing of software beamforming, demonstrated in this thesis, is that handheld devices, such as an iPhone, are powerful enough to implement the full ultrasound software processing chain. Software beamforming can also be considered a curse. A curse since the flexibility leads to methods that invalidate the metrics used to evaluate the image quality. This thesis evaluates the effects modern beamformers have on the ultrasound images, especially how adaptive beamformers influence, and invalidate the conventional contrast metrics. We introduce an improved contrast metric, the Generalized Contrast-to-Noise Ratio, immune to some of the unwanted effects.