Hand gesture recognition for doors with neural network

Hyun Jin Ahn, Jun Sung Kim, Jae Youn Shim, Jin Suk Kim

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

In this paper we propose a hand gesture recognition system for door opening. Because the usage of door knobs and the way of opening doors are similar worldwide, people will naturally do similar actions without special promise when opening the door. When a user wears a smart watch, it is possible to perform movements more natural than the movement at the situation with holding a smartphone in hand. We used an accelerometer embedded in a smart watch to collect hand gesture data, which opens each of three types of door, hinged, slide, and shutter. We preprocessed the raw data with two steps. We trimmed the data and normalized trimmed data using akima spline for multi-layer perceptron (MLP). Also, we usedMLP to classify the preprocessed hand gesture data in our system.

Original languageEnglish
Title of host publicationProceedings of the 2017 Research in Adaptive and Convergent Systems, RACS 2017
PublisherAssociation for Computing Machinery, Inc
Pages15-18
Number of pages4
ISBN (Electronic)9781450350273
DOIs
StatePublished - 20 Sep 2017
Event2017 International Conference on Research in Adaptive and Convergent Systems, RACS 2017 - Krakow, Poland
Duration: 20 Sep 201723 Sep 2017

Publication series

NameProceedings of the 2017 Research in Adaptive and Convergent Systems, RACS 2017
Volume2017-January

Conference

Conference2017 International Conference on Research in Adaptive and Convergent Systems, RACS 2017
Country/TerritoryPoland
CityKrakow
Period20/09/1723/09/17

Keywords

  • Deep Neural Network
  • Hand Gesture Recognition
  • Internet of Things
  • Smart Watch
  • Wearable Computing

Fingerprint

Dive into the research topics of 'Hand gesture recognition for doors with neural network'. Together they form a unique fingerprint.

Cite this