Intra-/inter-user adaptation framework for wearable gesture sensing device

Kosuke Kikui, Yuta Itoh, Makoto Yamada, Yuta Sugiura, Maki Sugimoto

研究成果: Conference contribution

1 引用 (Scopus)

抄録

The photo reflective sensor (PRS), a tiny distant-measurement module, is a popular electronic component widely used in wearable user-interfaces. An unavoidable issue of such wearable PRS devices in practical use is the need of user-independent training to have high gesture recognition accuracy. Each new user has to re-train a device by providing new training data (we call the inter-user setup). Even worse, re-training is also necessary ideally every time when the same user re-wears the device (we call the intra-user setup). In this paper, we propose a domain adaptation framework to reduce this training cost of users. Specifically, we adapt a pre-trained convolutional neural network (CNN) for both inter-user and intra-user setups to maintain the recognition accuracy high. We demonstrate, with an actual PRS device, that our framework significantly improves the average classification accuracy of the intra-user and inter-user setups up to 87.43% and 80.06% against the baseline (non-adapted) setups with the accuracy 68.96% and 63.26% respectively.

元の言語English
ホスト出版物のタイトルISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers
出版者Association for Computing Machinery
ページ21-24
ページ数4
ISBN(電子版)9781450359672
DOI
出版物ステータスPublished - 2018 10 8
イベント22nd International Symposium on Wearable Computers, ISWC 2018 - Singapore, Singapore
継続期間: 2018 10 82018 10 12

Other

Other22nd International Symposium on Wearable Computers, ISWC 2018
Singapore
Singapore
期間18/10/818/10/12

Fingerprint

Sensors
Gesture recognition
User interfaces
Wear of materials
Neural networks
Costs

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

これを引用

Kikui, K., Itoh, Y., Yamada, M., Sugiura, Y., & Sugimoto, M. (2018). Intra-/inter-user adaptation framework for wearable gesture sensing device. : ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers (pp. 21-24). Association for Computing Machinery. https://doi.org/10.1145/3267242.3267256

Intra-/inter-user adaptation framework for wearable gesture sensing device. / Kikui, Kosuke; Itoh, Yuta; Yamada, Makoto; Sugiura, Yuta; Sugimoto, Maki.

ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers. Association for Computing Machinery, 2018. p. 21-24.

研究成果: Conference contribution

Kikui, K, Itoh, Y, Yamada, M, Sugiura, Y & Sugimoto, M 2018, Intra-/inter-user adaptation framework for wearable gesture sensing device. : ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers. Association for Computing Machinery, pp. 21-24, 22nd International Symposium on Wearable Computers, ISWC 2018, Singapore, Singapore, 18/10/8. https://doi.org/10.1145/3267242.3267256
Kikui K, Itoh Y, Yamada M, Sugiura Y, Sugimoto M. Intra-/inter-user adaptation framework for wearable gesture sensing device. : ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers. Association for Computing Machinery. 2018. p. 21-24 https://doi.org/10.1145/3267242.3267256
Kikui, Kosuke ; Itoh, Yuta ; Yamada, Makoto ; Sugiura, Yuta ; Sugimoto, Maki. / Intra-/inter-user adaptation framework for wearable gesture sensing device. ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers. Association for Computing Machinery, 2018. pp. 21-24
@inproceedings{dbb0773d36d445d4afca391cab04695b,
title = "Intra-/inter-user adaptation framework for wearable gesture sensing device",
abstract = "The photo reflective sensor (PRS), a tiny distant-measurement module, is a popular electronic component widely used in wearable user-interfaces. An unavoidable issue of such wearable PRS devices in practical use is the need of user-independent training to have high gesture recognition accuracy. Each new user has to re-train a device by providing new training data (we call the inter-user setup). Even worse, re-training is also necessary ideally every time when the same user re-wears the device (we call the intra-user setup). In this paper, we propose a domain adaptation framework to reduce this training cost of users. Specifically, we adapt a pre-trained convolutional neural network (CNN) for both inter-user and intra-user setups to maintain the recognition accuracy high. We demonstrate, with an actual PRS device, that our framework significantly improves the average classification accuracy of the intra-user and inter-user setups up to 87.43{\%} and 80.06{\%} against the baseline (non-adapted) setups with the accuracy 68.96{\%} and 63.26{\%} respectively.",
keywords = "CNN, Domain Adaptation, Photo Reflective Sensor",
author = "Kosuke Kikui and Yuta Itoh and Makoto Yamada and Yuta Sugiura and Maki Sugimoto",
year = "2018",
month = "10",
day = "8",
doi = "10.1145/3267242.3267256",
language = "English",
pages = "21--24",
booktitle = "ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Intra-/inter-user adaptation framework for wearable gesture sensing device

AU - Kikui, Kosuke

AU - Itoh, Yuta

AU - Yamada, Makoto

AU - Sugiura, Yuta

AU - Sugimoto, Maki

PY - 2018/10/8

Y1 - 2018/10/8

N2 - The photo reflective sensor (PRS), a tiny distant-measurement module, is a popular electronic component widely used in wearable user-interfaces. An unavoidable issue of such wearable PRS devices in practical use is the need of user-independent training to have high gesture recognition accuracy. Each new user has to re-train a device by providing new training data (we call the inter-user setup). Even worse, re-training is also necessary ideally every time when the same user re-wears the device (we call the intra-user setup). In this paper, we propose a domain adaptation framework to reduce this training cost of users. Specifically, we adapt a pre-trained convolutional neural network (CNN) for both inter-user and intra-user setups to maintain the recognition accuracy high. We demonstrate, with an actual PRS device, that our framework significantly improves the average classification accuracy of the intra-user and inter-user setups up to 87.43% and 80.06% against the baseline (non-adapted) setups with the accuracy 68.96% and 63.26% respectively.

AB - The photo reflective sensor (PRS), a tiny distant-measurement module, is a popular electronic component widely used in wearable user-interfaces. An unavoidable issue of such wearable PRS devices in practical use is the need of user-independent training to have high gesture recognition accuracy. Each new user has to re-train a device by providing new training data (we call the inter-user setup). Even worse, re-training is also necessary ideally every time when the same user re-wears the device (we call the intra-user setup). In this paper, we propose a domain adaptation framework to reduce this training cost of users. Specifically, we adapt a pre-trained convolutional neural network (CNN) for both inter-user and intra-user setups to maintain the recognition accuracy high. We demonstrate, with an actual PRS device, that our framework significantly improves the average classification accuracy of the intra-user and inter-user setups up to 87.43% and 80.06% against the baseline (non-adapted) setups with the accuracy 68.96% and 63.26% respectively.

KW - CNN

KW - Domain Adaptation

KW - Photo Reflective Sensor

UR - http://www.scopus.com/inward/record.url?scp=85056814597&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056814597&partnerID=8YFLogxK

U2 - 10.1145/3267242.3267256

DO - 10.1145/3267242.3267256

M3 - Conference contribution

AN - SCOPUS:85056814597

SP - 21

EP - 24

BT - ISWC 2018 - Proceedings of the 2018 ACM International Symposium on Wearable Computers

PB - Association for Computing Machinery

ER -