JGU Tracking of Facial Movements

Abstract

The advent of modern neurophysiological imaging methods allow recordings of neural activity in a high temporal resolution. Compared to that, tools for quantifying behavior in a non-invasive manner lag behind, especially in terms of flexibility. In this project, we applied deep neural network-based pose estimation models  for the detection of facial movements in videotaped psychophysical experiments with rodents. The aim of this project is to explain variance in classical response data by means of video analysis.