Mobile-Optimized Real-Time Vessel Detection for Ultrasound-Guided Surgical Procedures

Non-compressible torso hemorrhage is the leading cause of potentially survivable fatalities in civilian and battlefield trauma. An insufficient number of trauma surgeons are expected to be available in future large-scale combat operations and natural disasters, creating a need for assistive technology to enable fast and accurate vascular access in pre-hospital environments. AI-GUIDE is a handheld surgical tool designed for emergency medical operations that combines an ultrasound probe with real-time image processing software, which controls robotic needle insertion components. Currently, AI-GUIDE relies on an external display/computer to perform the required computations. To reduce its size and weight, we investigate optimizations for mobile inference of the AI algorithms used in the AI-GUIDE prototype. Key optimizations include the use of mobile-optimized neural network models, quantization techniques, and leveraging a software development kit to harness portable hardware accel-eration. We compare the tradeoffs between speed and accuracy for different runtime configurations, quantization methods, and model sizes for two resource-conscious neural networks. We run our experiments on a smartphone as a reliable proxy for performance on the AI-GUIDE.