A low-cost freehand ultrasound probe tracking system using multimodal sensor fusion for 3D volume reconstruction

Freehand 3D ultrasound imaging requires accurate, real-time tracking of the probe’s position and orientation during acquisition to provide a comprehensive visualization of anatomical regions of interest. However, existing commercial tracking systems are often costly, and performance is subject to environmental constraints, such as optical occlusions and electromagnetic interference. In this work, we present a low-cost and robust tracking system that combines camera-based tracking with an inertial measurement sensor (IMU) embedded in a commercial wireless ultrasound probe for reliable motion tracking in resource-limited settings. We propose an Extended Kalman Filter (EKF)-based method to fuse the vision and IMU data, enabling robust estimation of the probe’s pose during freehand scanning, which is crucial for 3D volume reconstruction. The proposed approach was validated through phantom experiments. The tracking accuracy achieved for position and orientation was 2.305 ± 4.255mm and 0.174 ± 0.123°, respectively. Furthermore, the volumetric reconstruction using freehand 2D ultrasound acquisitions achieved a voxel-wise mean squared error (MSE) of 0.1365 and a dice similarity coefficient (DSC) of 0.8082, indicating a strong agreement with the ground truth volume. These results demonstrate the potential for combining low-cost camera and IMU systems to enhance tracking accuracy and improve clinical applications for 3D freehand ultrasound in resource-limited settings.