Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks

Eunsuk Chong, Taejin Choi, Hyungmin Kim, Seung-Jong Kim, Yoha Hwang, Jong Min Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We propose a novel approach of selecting useful input sensors as well as learning a mathematical model for predicting lower limb joint kinematics. We applied a feature selection method based on the mutual information called the variational information maximization, which has been reported as the state-of-the-art work among information based feature selection methods. The main difficulty in applying the method is estimating reliable probability density of input and output data, especially when the data are high dimensional and real-valued. We addressed this problem by applying a generative stochastic neural network called the restricted Boltzmann machine, through which we could perform sampling based probability estimation. The mutual informations between inputs and outputs are evaluated in each backward sensor elimination step, and the least informative sensor is removed with its network connections. The entire network is fine-tuned by maximizing conditional likelihood in each step. Experimental results are shown for 4 healthy subjects walking with various speeds, recording 64 sensor measurements including electromyogram, acceleration, and foot-pressure sensors attached on both lower limbs for predicting hip and knee joint angles. For test set of walking with arbitrary speed, our results show that our suggested method can select informative sensors while maintaining a good prediction accuracy.

Original languageEnglish
Title of host publication2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society
Subtitle of host publicationSmarter Technology for a Healthier World, EMBC 2017 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2043-2046
Number of pages4
ISBN (Electronic)9781509028092
DOIs
Publication statusPublished - 2017 Sep 13
Externally publishedYes
Event39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2017 - Jeju Island, Korea, Republic of
Duration: 2017 Jul 112017 Jul 15

Publication series

NameProceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
ISSN (Print)1557-170X

Other

Other39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2017
CountryKorea, Republic of
CityJeju Island
Period17/7/1117/7/15

Fingerprint

Biomechanical Phenomena
Lower Extremity
Kinematics
Learning
Neural networks
Sensors
Walking
Feature extraction
Hip Joint
Electromyography
Knee Joint
Foot
Pressure sensors
Healthy Volunteers
Theoretical Models
Joints
Pressure
Mathematical models
Sampling

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Cite this

Chong, E., Choi, T., Kim, H., Kim, S-J., Hwang, Y., & Lee, J. M. (2017). Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks. In 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Smarter Technology for a Healthier World, EMBC 2017 - Proceedings (pp. 2043-2046). [8037254] (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EMBC.2017.8037254

Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks. / Chong, Eunsuk; Choi, Taejin; Kim, Hyungmin; Kim, Seung-Jong; Hwang, Yoha; Lee, Jong Min.

2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Smarter Technology for a Healthier World, EMBC 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. p. 2043-2046 8037254 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chong, E, Choi, T, Kim, H, Kim, S-J, Hwang, Y & Lee, JM 2017, Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks. in 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Smarter Technology for a Healthier World, EMBC 2017 - Proceedings., 8037254, Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, Institute of Electrical and Electronics Engineers Inc., pp. 2043-2046, 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2017, Jeju Island, Korea, Republic of, 17/7/11. https://doi.org/10.1109/EMBC.2017.8037254
Chong E, Choi T, Kim H, Kim S-J, Hwang Y, Lee JM. Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks. In 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Smarter Technology for a Healthier World, EMBC 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2017. p. 2043-2046. 8037254. (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). https://doi.org/10.1109/EMBC.2017.8037254
Chong, Eunsuk ; Choi, Taejin ; Kim, Hyungmin ; Kim, Seung-Jong ; Hwang, Yoha ; Lee, Jong Min. / Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks. 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Smarter Technology for a Healthier World, EMBC 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 2043-2046 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).
@inproceedings{4909643dce5944ad86b76d31ce0b82e2,
title = "Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks",
abstract = "We propose a novel approach of selecting useful input sensors as well as learning a mathematical model for predicting lower limb joint kinematics. We applied a feature selection method based on the mutual information called the variational information maximization, which has been reported as the state-of-the-art work among information based feature selection methods. The main difficulty in applying the method is estimating reliable probability density of input and output data, especially when the data are high dimensional and real-valued. We addressed this problem by applying a generative stochastic neural network called the restricted Boltzmann machine, through which we could perform sampling based probability estimation. The mutual informations between inputs and outputs are evaluated in each backward sensor elimination step, and the least informative sensor is removed with its network connections. The entire network is fine-tuned by maximizing conditional likelihood in each step. Experimental results are shown for 4 healthy subjects walking with various speeds, recording 64 sensor measurements including electromyogram, acceleration, and foot-pressure sensors attached on both lower limbs for predicting hip and knee joint angles. For test set of walking with arbitrary speed, our results show that our suggested method can select informative sensors while maintaining a good prediction accuracy.",
author = "Eunsuk Chong and Taejin Choi and Hyungmin Kim and Seung-Jong Kim and Yoha Hwang and Lee, {Jong Min}",
year = "2017",
month = "9",
day = "13",
doi = "10.1109/EMBC.2017.8037254",
language = "English",
series = "Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "2043--2046",
booktitle = "2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society",

}

TY - GEN

T1 - Informative sensor selection and learning for prediction of lower limb kinematics using generative stochastic neural networks

AU - Chong, Eunsuk

AU - Choi, Taejin

AU - Kim, Hyungmin

AU - Kim, Seung-Jong

AU - Hwang, Yoha

AU - Lee, Jong Min

PY - 2017/9/13

Y1 - 2017/9/13

N2 - We propose a novel approach of selecting useful input sensors as well as learning a mathematical model for predicting lower limb joint kinematics. We applied a feature selection method based on the mutual information called the variational information maximization, which has been reported as the state-of-the-art work among information based feature selection methods. The main difficulty in applying the method is estimating reliable probability density of input and output data, especially when the data are high dimensional and real-valued. We addressed this problem by applying a generative stochastic neural network called the restricted Boltzmann machine, through which we could perform sampling based probability estimation. The mutual informations between inputs and outputs are evaluated in each backward sensor elimination step, and the least informative sensor is removed with its network connections. The entire network is fine-tuned by maximizing conditional likelihood in each step. Experimental results are shown for 4 healthy subjects walking with various speeds, recording 64 sensor measurements including electromyogram, acceleration, and foot-pressure sensors attached on both lower limbs for predicting hip and knee joint angles. For test set of walking with arbitrary speed, our results show that our suggested method can select informative sensors while maintaining a good prediction accuracy.

AB - We propose a novel approach of selecting useful input sensors as well as learning a mathematical model for predicting lower limb joint kinematics. We applied a feature selection method based on the mutual information called the variational information maximization, which has been reported as the state-of-the-art work among information based feature selection methods. The main difficulty in applying the method is estimating reliable probability density of input and output data, especially when the data are high dimensional and real-valued. We addressed this problem by applying a generative stochastic neural network called the restricted Boltzmann machine, through which we could perform sampling based probability estimation. The mutual informations between inputs and outputs are evaluated in each backward sensor elimination step, and the least informative sensor is removed with its network connections. The entire network is fine-tuned by maximizing conditional likelihood in each step. Experimental results are shown for 4 healthy subjects walking with various speeds, recording 64 sensor measurements including electromyogram, acceleration, and foot-pressure sensors attached on both lower limbs for predicting hip and knee joint angles. For test set of walking with arbitrary speed, our results show that our suggested method can select informative sensors while maintaining a good prediction accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85032200809&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85032200809&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2017.8037254

DO - 10.1109/EMBC.2017.8037254

M3 - Conference contribution

T3 - Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS

SP - 2043

EP - 2046

BT - 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society

PB - Institute of Electrical and Electronics Engineers Inc.

ER -