Design and 3-D applications of a new multi-channel ultrasound research system

The interaction between ultrasound waves and human body was initially exploited for therapeutic purposes, i.e. tissues were heated to treat some diseases. At the beginning of the 1950s, some scientists started exploiting ultrasound for generating images of the human body, thus enabling for diagnostic applications. In the last decades, ultrasound imaging considerably evolved and nowadays is one of the most used tools in clinical routine. However, research on ultrasound technologies has never rested and is, more than ever, very active worldwide. Ultrasound imaging involves different research fields, from the transducer technologies to the design of multi-channel systems, passing through signal processing techniques dedicated to beamforming, image enhancement, and artificial intelligence. The number of applications keeps growing and now includes 3-D imaging, blood flow estimations, tissue characterization, super-resolution, and more. The development and testing of new methods is usually based on the use of research scanners, or open platforms, i.e. ultrasound systems characterized by highly flexible hardware and software, which allow for a custom processing of the raw data acquired by the probe. The increasingly strong trend towards High-Frame-Rate (HFR) techniques together with the spread of matrix probes composed of thousands of elements, originated the need for research scanner architectures able to control such elements with an equal number of transmit/receive channels and manage the huge amount of received echo data according to HFR needs. This is a complex problem, considering that most open scanners, including the 256-channel Ultrasound Advanced Open Platform (ULA-OP 256), a research platform developed by the University of Florence, are limited to 256 channels. This Ph.D. work aimed at designing a new system architecture, composed of multiple ULA-OP 256 scanners, able to control a high number of channels and process the acquired data in real-time. The topology of the ULA-OP 256 was modified to improve its performance in 3-D applications in which a huge amount of data acquired using a 2-D probe must be transferred and processed at HFR. The main bottlenecks of ULA-OP 256 affecting the data transfer rate were detected and overtaken by optimizing the transfer of baseband data among the scanner boards, and by reducing the sample resolution of beamformed data. A master-slave architecture was designed to synchronize multiple ULA-OP 256 scanners and to control them as a single unified system, with high processing power and capable of controlling an arbitrary number of independent channels. The novel architecture was used to propose and test resource demanding methods that could not be implemented with a single ULA-OP 256. An HFR triplex bi-plane method was tested using two synchronized ULA-OP 256 connected to a 512-element sparse array probe, composed of two concentric 256-element sparse arrays. The new architecture allowed reconstructing at HFR both morphological and hemodynamic images, over two orthogonal planes. In collaboration with King’s College London, the same dual-scanner architecture was exploited to extend and test the Coherent Multi-Transducer UltraSound (CoMTUS) method to 3-D . Also, the same setup was used to propose and test a dual-probe HFR vector Doppler technique for accurate and repeatable estimations of the 2-D blood velocity vectors over a 3-D Region Of Interest (ROI). Finally, exploiting the real-time processing power of the two scanners, a 2-D CoMTUS modality was run in real-time for the first time.