Freehand Ultrasound

Freehand ultrasound aims to reconstruct three-dimensional images from two-dimensional scans acquired with a handheld probe, eliminating the need for bulky tracking systems. Current research heavily utilizes deep learning, employing various neural network architectures like recurrent networks and convolutional neural networks, often incorporating inertial measurement units (IMUs) for improved accuracy and robustness. This technology promises to improve accessibility and affordability of 3D ultrasound imaging, particularly in applications like cardiac echocardiography and neurosurgery, by enabling more precise and efficient image acquisition and analysis.