Peruvian sign language recognition using low resolution cameras

Bryan Berru-Novoa, Ricardo Gonzalez-Valenzuela, Pedro Shiguihara-Juarez

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

6 Scopus citations

Abstract

The recognition of sign language gesture through image processing and Machine Learning has been widely studied in recent years. This article presents a dataset consisting of 2400 images of the static gestures of the Peruvian sign language alphabet, in addition to applying it to a hand gesture recognition system using low resolution cameras. For the gesture recognition, the Histogram Oriented Gradient feature descriptor was used, along with 4 classification algorithms. The results showed that Histogram Oriented Gradient, along with Support Vector Machine, got the best result with a 89.46% accuracy and the system was able to recognize the gestures with variations of translation, rotation and scale.

Original languageEnglish
Title of host publicationProceedings of the 2018 IEEE 25th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538654903
DOIs
StatePublished - 6 Nov 2018
Externally publishedYes
Event25th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018 - Lima, Peru
Duration: 8 Aug 201810 Aug 2018

Publication series

NameProceedings of the 2018 IEEE 25th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018

Conference

Conference25th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2018
Country/TerritoryPeru
CityLima
Period8/08/1810/08/18

Keywords

  • Histogram Oriented Gradient
  • Image Processing
  • Machine Learning
  • Sign Language Recognition

Fingerprint

Dive into the research topics of 'Peruvian sign language recognition using low resolution cameras'. Together they form a unique fingerprint.

Cite this