On the use of an extended Kalman filter for subviral particle tracking

Abstract

The ongoing outbreaks of the Ebola disease in the Democratic Republic of Congo demand a faster medical reaction. To accelerate the search for an antiviral, processes need to be automated. Previously, algorithms to automatically detect, track and analyze subviral particles in fluoroscopic image sequences were presented. Thereby, a linear Kalman filter algorithm is used to improve the tracking. In this publication the predictions of the linear and an extended Kalman filter are compared. Both approaches are tested on a real subviral particle track and show that an extended Kalman filter is suitable for the complex motion patterns of subviral particles.