A signal-domain method to reconstruct object’s boundary for 3D ultrasound computed tomography

3D Ultrasound Computed Tomography (USCT) is inherently computationally intensive, primarily because a large imaging volume including the object (e.g., human breast) and coupling medium needs to be reconstructed. Knowing the boundary of the object prior to image reconstruction could aid in reducing the required computational resources. In this work, we propose a signal-domain object segmentation method to retrieve the 3D boundary of the object prior to image reconstruction. Using the experimental data of a phantom, the performance of the reconstruction algorithm is evaluated.