Real-Time Quantitative Ultrasound and Radar Medical Imaging

Medical imaging is crucial for diagnosing, monitoring, and treating medical conditions. Ultrasound and radar signals are useful for medical imaging due to their non-invasive, non-ionizing, low-cost, and accessible nature. However, traditional imaging techniques lack physical interpretation. Quantitative medical imaging enables a more comprehensive understanding of physical properties and is beneficial for broader range of applications such as cancer detection, improving fatty liver diagnosis, and stroke imaging. Current quantitative techniques such as Full Waveform Inversion suffer from time-consuming processes and a tendency to converge to local minima, which can lead to unsatisfactory outcomes. To overcome these challenges we present a neural network that considers in its design the symmetries and properties of the received signals to enable real-time quantitative mappings of physical properties. Our proposed method attains a high level of accuracy for complex shapes with less than 0.15 seconds per test sample, compared to 0.75-2 hours for the competing method.