Ultrasound Beamforming: FPGA Implementation and Dynamic Focusing

Abstract

Beamformer is a core component in ultrasound systems, which stands as a signal processing unit forming a beam in a desired direction from interfering signals at each channel by adjusting their weights and delays. Upon reception by the transducer, the signals are sampled using high speed analog to digital converters. These samples are then beamformed using a delay profile generated for a single focal point. Implementing the beamformer in hardware for efficient frame rates, typically required in medical applications, is a crucial task involving handling of data sampled at high sampling frequency, parallel processing and inter-medium bandwidth constraints. This thesis investigates the implementation of a fixed focus delay and sum beamformer on the Ultrasound Array Research Platform (UARP) configured at University of Leeds, UK. Furthermore, fixed focus beamforming often results in poor image quality with focus fixed at one pixel location only. Dynamic focusing is an alternative approach where every pixel in the image is focused. This research also presents a study of the dynamic focus delay and sum beamformer to enhance the image quality and signal to noise (SNR) ratio. In this work, feasibility study is performed, and architecture is proposed for performing beamforming in real time. Such architecture can be implemented on FPGAs and can be fabricated into ASICs. The proposed system can be installed with any Ultrasound system being modular; hence it removed the platform dependency. This work resulted in implementation of a real time FPGA based delay and sum beamformer for Ultrasound system (UARP). The hardware replacement of the software based beamformer in UARP increased its frame rate to about 10 times while further optimization provided frame rate as high as 32 frames per second as compared to only 1 frame per second initially. Also, the similar design was studied for dynamic focusing instead of fixed focusing and it was observed that the signal to noise ratio (SNR) increased as compared to the fixed focus beamforming. The image resolution also increased which enhances the system output in terms of image quality.