Development of magnetic-sensor-based hand gesture recognition system for sign language

Bo Shi, Xi Chen*, Zhongzheng He, Ruoyu Han

*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

Sign language recognition is essential for the automatic translation of sign languages to enable communication for hearing-impaired people. This work proposes a system based on multiple magnetic sensors for recognizing hand gestures related to sign language alphabets. In particular, a magnetic detection system consisting of six magnetic sensor nodes measures the orientation of fingers and palms. A deep learning classification algorithm processes the measured orientation data. Experimental tests validate the proposed system and classification method. The results show that the proposed method provides close to 100% classification accuracy for 26 sign language alphabets under laboratory conditions. Thus, the feasibility of the proposed gesture recognition system for automatic translation of sign language alphabets is proved.

Original languageEnglish
Title of host publication2023 IEEE 6th International Electrical and Energy Conference, CIEEC 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2302-2305
Number of pages4
ISBN (Electronic)9798350346671
DOIs
Publication statusPublished - 2023
Event6th IEEE International Electrical and Energy Conference, CIEEC 2023 - Hefei, China
Duration: 12 May 202314 May 2023

Publication series

Name2023 IEEE 6th International Electrical and Energy Conference, CIEEC 2023

Conference

Conference6th IEEE International Electrical and Energy Conference, CIEEC 2023
Country/TerritoryChina
CityHefei
Period12/05/2314/05/23

Keywords

  • IoT
  • gesture recognition
  • magnetic detection
  • status detection
  • system integration

Fingerprint

Dive into the research topics of 'Development of magnetic-sensor-based hand gesture recognition system for sign language'. Together they form a unique fingerprint.

Cite this