SBIR Phase I: AI-powered architecture for smarter ultrasound applications in healthcare

这个小企业创新研究（SBIR）第一阶段项目的广泛影响是通过引入超声系统的创新架构，在医学成像领域开辟新的领域。超声检查是一种广泛应用于临床诊断和指导治疗程序的方法。然而，超声波发动机的技术瓶颈目前阻碍了数据密集型应用。通过将计算能力直接引入超声引擎，该项目将实现具有最高质量的实时3D功能超声成像。提出的创新将使新一代超声设备能够以非常详细的方式可视化身体组织，并由数据驱动的计算机辅助诊断应用程序提供动力。超声系统将首次拥有足够高质量的数据来使用人工智能（AI）算法，这有望减少与超声分析相关的时间和成本，并改善患者的治疗效果。该项目还将促进超声技术在医学应用领域的研究进展，使新一代工具能够在世界各地的研究机构中使用。拟议的项目将改变超声发动机的范例。今天，超声数据密集型应用受到从超声引擎到计算机的信号传输速度的限制，在那里它们必须被处理。本项目提出在超声引擎内部集成专门的计算能力，在采集信号的地方进行实时信号处理和图像分析，降低对数据传输和处理的要求。第一阶段项目所建议的活动将允许公司构建和验证高性能可扩展硬件架构的功能原型。原型机执行超快速帧形成成像的能力将通过模拟数据密集型操作来验证。同时，将开发一个最小可行的软件来控制、分析和实时显示硬件产生的数据量，并动态加载临床使用的应用程序和算法。本一期项目的可行性研究将作为拓展软件平台运行能力的基础，并评估样机对人体结构进行实时高速成像和识别的能力。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact of this Small Business Innovation Research (SBIR) Phase I project is to open new frontiers in medical imaging by introducing an innovative architecture for ultrasound systems. Ultrasonography is a widely used approach to perform clinical diagnosis and guide therapeutic procedures. However, technological bottlenecks in ultrasound engines currently prevent data intensive applications. By bringing the computational power directly within the ultrasound engine, this project will allow 3D functional ultrasound imaging in real-time with the highest quality. The innovation proposed will allow the generation of new ultrasound devices able to visualize body tissues in exceptional details and powered with data-driven computer-aided diagnostic applications. For the first time, ultrasound systems will have enough quality data to use algorithms of artificial intelligence (AI), which are expected to reduce time and costs associated with ultrasound analysis as well as improve patient outcomes. This project will also contribute to the advancement of research in the field of medical applications of ultrasonography technology by allowing the generation of new tools to be used in Research Institutes worldwide. The proposed project will shift the paradigm in ultrasound engines. Today, ultrasound data-intense applications are restricted by the limitation in transmission speed of signals from the ultrasound engine to a computer where they have to be processed. This project proposes to integrate specialized computational power inside the ultrasound engine to allow real-time signal processing and image analysis where the signal is collected, hence reducing the requirements for the data transmission and processing. The activities proposed for this Phase I project will allow the company to build and validate a functioning prototype of high-performance scalable hardware architecture. The ability of the prototype to perform ultra-fast frame-forming imaging will be validated by simulating data-intensive operations. Concurrently, a minimal viable software will be developed to control, analyze and display in real-time the volume of data produced by the hardware, and to dynamically load applications and algorithms for clinical use. The feasibility study performed in this Phase I project will constitute the basis to expand the operation capabilities of the software platform and assess the ability of the prototype to perform real-time high speed imaging and recognition of bodily structures.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.