Unsupervised and Generic Short-Term Anticipation of Human Body Motions

Kristina Enes, Hassan Errami, Moritz Wolter, Tim Krake, Bernd Eberhardt, Andreas Weber und Jörg Zimmermann
In: Sensors (Feb. 2020), 20:4:976
 

Abstract

Various neural network based methods are capable of anticipating human body motions from data for a short period of time. What these methods lack are the interpretability and explainability of the network and its results. We propose to use Dynamic Mode Decomposition with delays to represent and anticipate human body motions. Exploring the influence of the number of delays on the reconstruction and prediction of various motion classes, we show that the anticipation errors in our results are comparable or even better for very short anticipation times (<0.4 sec) to a recurrent neural network based method. We perceive our method as a first step towards the interpretability of the results by representing human body motions as linear combinations of “factors”. In addition, compared to the neural network based methods large training times are not needed. Actually, our methods do not even regress to any other motions than the one to be anticipated and hence is of a generic nature.

Stichwörter: delay coordinates, dynamic mode decomposition, human motion anticipation, short-time future prediction

Bilder

Bibtex

@ARTICLE{eewkewz19,
    author = {Enes, Kristina and Errami, Hassan and Wolter, Moritz and Krake, Tim and Eberhardt, Bernd and Weber,
              Andreas and Zimmermann, J{\"o}rg},
     title = {Unsupervised and Generic Short-Term Anticipation of Human Body Motions},
   journal = {Sensors},
    volume = {20},
    number = {4:976},
      year = {2020},
     month = feb,
  keywords = {delay coordinates, dynamic mode decomposition, human motion anticipation, short-time future
              prediction},
  abstract = {Various neural network based methods are capable of anticipating human body motions from data for a
              short period of time. What these methods lack are the interpretability and explainability of the
              network and its results. We propose to use Dynamic Mode Decomposition with delays to represent and
              anticipate human body motions. Exploring the influence of the number of delays on the reconstruction
              and prediction of various motion classes, we show that the anticipation errors in our results are
              comparable or even better for very short anticipation times (<0.4 sec) to a recurrent neural network
              based method. We perceive our method as a first step towards the interpretability of the results by
              representing human body motions as linear combinations of ``factors''. In addition, compared to the
              neural network based methods large training times are not needed. Actually, our methods do not even
              regress to any other motions than the one to be anticipated and hence is of a generic nature.},
       url = {https://www.mdpi.com/1424-8220/20/4/976},
       doi = {10.3390/s20040976}
}