Collaborative classification for daily activity recognition with a smartwatch

Hyunchoong Kim, Jonghoon Shin, Soohwan Kim, Yohan Ko, Kyoungwoo Lee, Hojung Cha, Seong Il Hahm, Taejun Kwon

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

3 Citations (Scopus)

Abstract

Research of daily activity recognition has been extensively conducted in the field of ubiquitous computing. However, previous daily activity recognition schemes are either obtrusive or inaccurate since they use just special-purpose devices. In this paper, we propose the collaborative classification for recognizing daily activities with a smartwatch. We exploit a single off-the-shelf smartwatch to distinguish 5 different daily activities such as eating, vacuuming, sleeping, showering, and TV watching. More precisely, we conduct experiments for collecting sensor data from accelerometer and acoustic sensor which are embedded in a smartwatch. However, the simple combination of the raw acceleration and acoustic data does not deliver accurate recognition accuracy. In order to achieve high accuracy, we propose a collaborative classification algorithm which integrates sensor data and ground-truth label for improving recognition accuracy by constructing a mapping table. We evaluate accuracies using single-sensor based approach, multi-sensor based approach, and our collaborative classification approach. The results from activity recognition for about 20 hours data collected by subjects show reliable accuracies for all 5 activities, and the overall accuracy of our collaborative approach is about 91.5%. Experimental results reveal that our approach improves the recall rate of each activity by up to 21.5% as compared to that of the simply combined multi-sensor based approach.

Original languageEnglish
Title of host publication2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3707-3712
Number of pages6
ISBN (Electronic)9781509018970
DOIs
Publication statusPublished - 2017 Feb 6
Event2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Budapest, Hungary
Duration: 2016 Oct 92016 Oct 12

Publication series

Name2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings

Other

Other2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016
CountryHungary
CityBudapest
Period16/10/916/10/12

Fingerprint

Activity Recognition
Sensors
Sensor
Acoustics
Accelerometer
Ubiquitous Computing
Classification Algorithm
Inaccurate
Ubiquitous computing
Table
Accelerometers
High Accuracy
Integrate
Labels
Evaluate
Experimental Results
Experiment
Experiments

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition
  • Artificial Intelligence
  • Control and Optimization
  • Human-Computer Interaction

Cite this

Kim, H., Shin, J., Kim, S., Ko, Y., Lee, K., Cha, H., ... Kwon, T. (2017). Collaborative classification for daily activity recognition with a smartwatch. In 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings (pp. 3707-3712). [7844810] (2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SMC.2016.7844810
Kim, Hyunchoong ; Shin, Jonghoon ; Kim, Soohwan ; Ko, Yohan ; Lee, Kyoungwoo ; Cha, Hojung ; Hahm, Seong Il ; Kwon, Taejun. / Collaborative classification for daily activity recognition with a smartwatch. 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 3707-3712 (2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings).
@inproceedings{d67b6b001daf442bb7fd4471faa50cbd,
title = "Collaborative classification for daily activity recognition with a smartwatch",
abstract = "Research of daily activity recognition has been extensively conducted in the field of ubiquitous computing. However, previous daily activity recognition schemes are either obtrusive or inaccurate since they use just special-purpose devices. In this paper, we propose the collaborative classification for recognizing daily activities with a smartwatch. We exploit a single off-the-shelf smartwatch to distinguish 5 different daily activities such as eating, vacuuming, sleeping, showering, and TV watching. More precisely, we conduct experiments for collecting sensor data from accelerometer and acoustic sensor which are embedded in a smartwatch. However, the simple combination of the raw acceleration and acoustic data does not deliver accurate recognition accuracy. In order to achieve high accuracy, we propose a collaborative classification algorithm which integrates sensor data and ground-truth label for improving recognition accuracy by constructing a mapping table. We evaluate accuracies using single-sensor based approach, multi-sensor based approach, and our collaborative classification approach. The results from activity recognition for about 20 hours data collected by subjects show reliable accuracies for all 5 activities, and the overall accuracy of our collaborative approach is about 91.5{\%}. Experimental results reveal that our approach improves the recall rate of each activity by up to 21.5{\%} as compared to that of the simply combined multi-sensor based approach.",
author = "Hyunchoong Kim and Jonghoon Shin and Soohwan Kim and Yohan Ko and Kyoungwoo Lee and Hojung Cha and Hahm, {Seong Il} and Taejun Kwon",
year = "2017",
month = "2",
day = "6",
doi = "10.1109/SMC.2016.7844810",
language = "English",
series = "2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "3707--3712",
booktitle = "2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings",
address = "United States",

}

Kim, H, Shin, J, Kim, S, Ko, Y, Lee, K, Cha, H, Hahm, SI & Kwon, T 2017, Collaborative classification for daily activity recognition with a smartwatch. in 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings., 7844810, 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 3707-3712, 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016, Budapest, Hungary, 16/10/9. https://doi.org/10.1109/SMC.2016.7844810

Collaborative classification for daily activity recognition with a smartwatch. / Kim, Hyunchoong; Shin, Jonghoon; Kim, Soohwan; Ko, Yohan; Lee, Kyoungwoo; Cha, Hojung; Hahm, Seong Il; Kwon, Taejun.

2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. p. 3707-3712 7844810 (2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings).

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

TY - GEN

T1 - Collaborative classification for daily activity recognition with a smartwatch

AU - Kim, Hyunchoong

AU - Shin, Jonghoon

AU - Kim, Soohwan

AU - Ko, Yohan

AU - Lee, Kyoungwoo

AU - Cha, Hojung

AU - Hahm, Seong Il

AU - Kwon, Taejun

PY - 2017/2/6

Y1 - 2017/2/6

N2 - Research of daily activity recognition has been extensively conducted in the field of ubiquitous computing. However, previous daily activity recognition schemes are either obtrusive or inaccurate since they use just special-purpose devices. In this paper, we propose the collaborative classification for recognizing daily activities with a smartwatch. We exploit a single off-the-shelf smartwatch to distinguish 5 different daily activities such as eating, vacuuming, sleeping, showering, and TV watching. More precisely, we conduct experiments for collecting sensor data from accelerometer and acoustic sensor which are embedded in a smartwatch. However, the simple combination of the raw acceleration and acoustic data does not deliver accurate recognition accuracy. In order to achieve high accuracy, we propose a collaborative classification algorithm which integrates sensor data and ground-truth label for improving recognition accuracy by constructing a mapping table. We evaluate accuracies using single-sensor based approach, multi-sensor based approach, and our collaborative classification approach. The results from activity recognition for about 20 hours data collected by subjects show reliable accuracies for all 5 activities, and the overall accuracy of our collaborative approach is about 91.5%. Experimental results reveal that our approach improves the recall rate of each activity by up to 21.5% as compared to that of the simply combined multi-sensor based approach.

AB - Research of daily activity recognition has been extensively conducted in the field of ubiquitous computing. However, previous daily activity recognition schemes are either obtrusive or inaccurate since they use just special-purpose devices. In this paper, we propose the collaborative classification for recognizing daily activities with a smartwatch. We exploit a single off-the-shelf smartwatch to distinguish 5 different daily activities such as eating, vacuuming, sleeping, showering, and TV watching. More precisely, we conduct experiments for collecting sensor data from accelerometer and acoustic sensor which are embedded in a smartwatch. However, the simple combination of the raw acceleration and acoustic data does not deliver accurate recognition accuracy. In order to achieve high accuracy, we propose a collaborative classification algorithm which integrates sensor data and ground-truth label for improving recognition accuracy by constructing a mapping table. We evaluate accuracies using single-sensor based approach, multi-sensor based approach, and our collaborative classification approach. The results from activity recognition for about 20 hours data collected by subjects show reliable accuracies for all 5 activities, and the overall accuracy of our collaborative approach is about 91.5%. Experimental results reveal that our approach improves the recall rate of each activity by up to 21.5% as compared to that of the simply combined multi-sensor based approach.

UR - http://www.scopus.com/inward/record.url?scp=85015761708&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015761708&partnerID=8YFLogxK

U2 - 10.1109/SMC.2016.7844810

DO - 10.1109/SMC.2016.7844810

M3 - Conference contribution

AN - SCOPUS:85015761708

T3 - 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings

SP - 3707

EP - 3712

BT - 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Kim H, Shin J, Kim S, Ko Y, Lee K, Cha H et al. Collaborative classification for daily activity recognition with a smartwatch. In 2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings. Institute of Electrical and Electronics Engineers Inc. 2017. p. 3707-3712. 7844810. (2016 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2016 - Conference Proceedings). https://doi.org/10.1109/SMC.2016.7844810