Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks

Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks

Optical motion capture is a mature contemporary technique for the acquisition of motion data; alas, it is non-error-free. Due to technical limitations and occlusions of markers, gaps might occur in such recordings. The article reviews various neural network architectures applied to the gap-filling p...

Full description

Bibliographic Details
Main Authors: Przemysław Skurowski, Magdalena Pawlyta
Format: Article
Language:English
Published: MDPI AG 2021-09-01
Series:Sensors
Subjects:
motion capture
neural networks
reconstruction
gap filling
FFNN
LSTM
Online Access:https://www.mdpi.com/1424-8220/21/18/6115
id doaj-d5592f80c73a4f738045150bc243070b
record_format Article
spelling doaj-d5592f80c73a4f738045150bc243070b2021-09-26T01:22:45ZengMDPI AGSensors1424-82202021-09-01216115611510.3390/s21186115Gap Reconstruction in Optical Motion Capture Sequences Using Neural NetworksPrzemysław Skurowski0Magdalena Pawlyta1Department of Graphics, Computer Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, PolandDepartment of Graphics, Computer Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, PolandOptical motion capture is a mature contemporary technique for the acquisition of motion data; alas, it is non-error-free. Due to technical limitations and occlusions of markers, gaps might occur in such recordings. The article reviews various neural network architectures applied to the gap-filling problem in motion capture sequences within the FBM framework providing a representation of body kinematic structure. The results are compared with interpolation and matrix completion methods. We found out that, for longer sequences, simple linear feedforward neural networks can outperform the other, sophisticated architectures, but these outcomes might be affected by the small amount of data availabe for training. We were also able to identify that the acceleration and monotonicity of input sequence are the parameters that have a notable impact on the obtained results.https://www.mdpi.com/1424-8220/21/18/6115motion captureneural networksreconstructiongap fillingFFNNLSTM
collection DOAJ
language English
format Article
sources DOAJ
author Przemysław Skurowski
Magdalena Pawlyta
spellingShingle Przemysław Skurowski
Magdalena Pawlyta
Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
Sensors
motion capture
neural networks
reconstruction
gap filling
FFNN
LSTM
author_facet Przemysław Skurowski
Magdalena Pawlyta
author_sort Przemysław Skurowski
title Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
title_short Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
title_full Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
title_fullStr Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
title_full_unstemmed Gap Reconstruction in Optical Motion Capture Sequences Using Neural Networks
title_sort gap reconstruction in optical motion capture sequences using neural networks
publisher MDPI AG
series Sensors
issn 1424-8220
publishDate 2021-09-01
description Optical motion capture is a mature contemporary technique for the acquisition of motion data; alas, it is non-error-free. Due to technical limitations and occlusions of markers, gaps might occur in such recordings. The article reviews various neural network architectures applied to the gap-filling problem in motion capture sequences within the FBM framework providing a representation of body kinematic structure. The results are compared with interpolation and matrix completion methods. We found out that, for longer sequences, simple linear feedforward neural networks can outperform the other, sophisticated architectures, but these outcomes might be affected by the small amount of data availabe for training. We were also able to identify that the acceleration and monotonicity of input sequence are the parameters that have a notable impact on the obtained results.
topic motion capture
neural networks
reconstruction
gap filling
FFNN
LSTM
url https://www.mdpi.com/1424-8220/21/18/6115
work_keys_str_mv AT przemysławskurowski gapreconstructioninopticalmotioncapturesequencesusingneuralnetworks
AT magdalenapawlyta gapreconstructioninopticalmotioncapturesequencesusingneuralnetworks
_version_ 1716869103955738624

Similar Items