Title: Calibration Using Matrix Completion With Application to Ultrasound Tomography.

Authors: Reza Parhizkar, Amin Karbasi, Sewoong Oh, Martin Vetterli


We study the application of matrix completion in the process of calibrating physical devices. In particular we propose an algorithm together with reconstruction bounds for calibrating circular ultrasound tomography devices. We use the time-of-flight (ToF) measurements between sensor pairs in a homogeneous medium to calibrate the system. The calibration process consists of a low-rank matrix completion algorithm to de-noise and estimate random and structured missing ToFs, and the classic multi-dimensional scaling method to estimate the sensor positions from the ToF measurements. We provide theoretical bounds on the calibration error. Several simulations are conducted to evaluate the theoretical results presented in this paper.

Full Text: [PDF]

Accessibility at Yale   Inference, Information, and Decision Group at Yale