Motion correction of thermographic images in neurosurgery: Performance comparison

Abstract

In this paper, the correction of motion-induced high frequency artifacts on sequences of thermographic images by a real-time capable approach based on Cellular Nonlinear Networks (CNN) is evaluated. For comparison, an offline analysis of frequency subspaces is presented and results will be compared. Both simulated data sets as well as data sets recorded during neurosurgery are considered as inputs and two measures are evaluated.