Erratum: Human Movement Representation on Multivariate Time Series for Recognition of Professional Gestures and Forecasting Their Trajectories

An erratum on
Human Movement Representation on Multivariate Time Series for Recognition of Professional Gestures and Forecasting Their Trajectories

by Manitsaris, S., Senteri, G., Makrygiannis, D., and Glushkova, A. (2020). Front. Robot. AI 7:80. doi: 10.3389/frobt.2020.00080

Due to a production error, several words were incorrectly capitalized throughout the text. These have now been correctly capitalized.

In Equations 17–20, there were several instances where the authors mistakenly switched the positive and negative signs.

In the Statistical Significance and Simulation of the Models section, subsection Recognition Accuracy and Comparison and End-to-End Deep Learning Architectures, paragraph 12, the authors accidently missed the addition of a footnote. The following footnote has been added for UCF-101: https://www.crcv.ucf.edu/data/UCF101.php.

In Table 7, “Kingma and Ba (2014)” should be replaced with “Tran et al. (2015).” Following that, “Tran et al. (2015)” should be replaced with “Competé et al. (2019).”

In the legend for Figure 6, a sentence has been added to clarify that this is based on “G_2,4” where 2 & 4 are pointers.

In the Statistical Significance and Simulation of the Models section, subsection Forecasting Ability for Motion Trajectories, paragraph 2, The final two mentions of “G_2,1” have been corrected to “G_2,4.”

The publisher and authors apologize for these errors. The original version of this article has been updated.