Smart Percussion for automated diagnostics

这项研究的目的是开发一种智能撞击设备，以帮助医疗从业者更有效地诊断患者，减少对有害扫描和创伤干预的需求。该设备可以部署在初级和二级护理之外。开发该设备的关键要素是一个进行敲击实验的模型和两种不同的行为建模方法：通过求解声音传播的控制方程而开发的基于物理的模型，还有一台机器通过使用幻影观测训练分类器开发的基于学习的模型。该项目福尔斯数字经济和医疗保健技术EPSRC的主题和医疗保健技术的重大挑战。它将通过支持开发新设备来直接做出贡献，以提高诊断效率，最大限度地降低成本，同时降低患者风险。

The aim of this research is to develop a smart percussion device to help medical practitioners diagnose patients more efficiently, reducing the need for harmful scans and traumatising interventions. The device could be deployed outside of primary and secondary care.The key elements in the development of this device are a phantom on which to perform percussion experiments and two different approaches to modelling the behaviour: a physics-based model developed by solving the governing equation for sound propagation, and a machine-learning based model developed by training a classifier using phantom observations.The project falls under Digital Economies and Healthcare Technologies themes of EPSRC and under Healthcare Technologies Grand Challenges. It would contribute directly by supporting the development of novel devices to enhance the efficacy of diagnosis, minimising costs while reducing risk to patients.