We present a pipeline for recognizing dynamic freehand gestures on mobile devices based on extracting depth information coming from a single Time-of-Flight sensor. Hand gestures are recorded with a mobile 3D sensor, transformed frame by frame into an appropriate 3D descriptor and fed into a deep LSTM network for recognition purposes. LSTM being a recurrent neural model, it is uniquely suited for classifying explicitly time-dependent data such as hand gestures. For training and testing purposes, we create a small database of four hand gesture classes, each comprising 40 \$\$\backslashtimes \$\$× 150 3D frames. We conduct experiments concerning execution speed on a mobile device, generalization capability as a function of network topology, and classification ability `ahead of time', i.e., when the gesture is not yet completed. Recognition rates are high (>95\%) and maintainable in real-time as a single classification step requires less than 1 ms computation time, introducing freehand gestures for mobile systems.
Description
Dynamic Hand Gesture Recognition for Mobile Systems Using Deep LSTM | SpringerLink
%0 Conference Paper
%1 sarkar2017dynamic
%A Sarkar, Ayanava
%A Gepperth, Alexander
%A Handmann, Uwe
%A Kopinski, Thomas
%B Intelligent Human Computer Interaction
%C Cham
%D 2017
%E Horain, Patrick
%E Achard, Catherine
%E Mallem, Malik
%I Springer
%K deep-learning gestural-interaction mobile-computing real
%P 19-31
%R 10.1007/978-3-319-72038-8_3
%T Dynamic Hand Gesture Recognition for Mobile Systems Using Deep LSTM
%U https://doi.org/10.1007/978-3-319-72038-8_3
%V 10688
%X We present a pipeline for recognizing dynamic freehand gestures on mobile devices based on extracting depth information coming from a single Time-of-Flight sensor. Hand gestures are recorded with a mobile 3D sensor, transformed frame by frame into an appropriate 3D descriptor and fed into a deep LSTM network for recognition purposes. LSTM being a recurrent neural model, it is uniquely suited for classifying explicitly time-dependent data such as hand gestures. For training and testing purposes, we create a small database of four hand gesture classes, each comprising 40 \$\$\backslashtimes \$\$× 150 3D frames. We conduct experiments concerning execution speed on a mobile device, generalization capability as a function of network topology, and classification ability `ahead of time', i.e., when the gesture is not yet completed. Recognition rates are high (>95\%) and maintainable in real-time as a single classification step requires less than 1 ms computation time, introducing freehand gestures for mobile systems.
%@ 978-3-319-72038-8
@inproceedings{sarkar2017dynamic,
abstract = {We present a pipeline for recognizing dynamic freehand gestures on mobile devices based on extracting depth information coming from a single Time-of-Flight sensor. Hand gestures are recorded with a mobile 3D sensor, transformed frame by frame into an appropriate 3D descriptor and fed into a deep LSTM network for recognition purposes. LSTM being a recurrent neural model, it is uniquely suited for classifying explicitly time-dependent data such as hand gestures. For training and testing purposes, we create a small database of four hand gesture classes, each comprising 40 {\$}{\$}{\backslash}times {\$}{\$}{\texttimes} 150 3D frames. We conduct experiments concerning execution speed on a mobile device, generalization capability as a function of network topology, and classification ability `ahead of time', i.e., when the gesture is not yet completed. Recognition rates are high (>95{\%}) and maintainable in real-time as a single classification step requires less than 1 ms computation time, introducing freehand gestures for mobile systems.},
added-at = {2019-11-14T07:01:57.000+0100},
address = {Cham},
author = {Sarkar, Ayanava and Gepperth, Alexander and Handmann, Uwe and Kopinski, Thomas},
biburl = {https://www.bibsonomy.org/bibtex/24c8b54eacafbc0f37b2f1b8b146b79e5/jpmor},
booktitle = {Intelligent Human Computer Interaction},
description = {Dynamic Hand Gesture Recognition for Mobile Systems Using Deep LSTM | SpringerLink},
doi = {10.1007/978-3-319-72038-8_3},
editor = {Horain, Patrick and Achard, Catherine and Mallem, Malik},
interhash = {3602ed6f28f25dae923e62742330a745},
intrahash = {4c8b54eacafbc0f37b2f1b8b146b79e5},
isbn = {978-3-319-72038-8},
keywords = {deep-learning gestural-interaction mobile-computing real},
language = {English},
pages = {19-31},
publisher = {Springer},
school = {Birla Institute of Technology and Science (BITS)},
series = {Lecture Notes in Computer Science},
timestamp = {2020-10-07T13:36:50.000+0200},
title = {Dynamic Hand Gesture Recognition for Mobile Systems Using Deep LSTM},
url = {https://doi.org/10.1007/978-3-319-72038-8_3},
volume = 10688,
year = 2017
}