Design and development of sign language questionnaires based on video and web interfaces

Juan Pedro López*, Marta Bosch-Baliarda, Carlos Alberto Martín, José Manuel Menéndez, Pilar Orero, Olga Soler, Federico Álvarez

*Corresponding author for this work

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Conventional tests with written information used for the evaluation of sign language (SL) comprehension introduce distortions due to the translation process. This fact affects the results and conclusions drawn and, for that reason, it is necessary to design and implement the same language interpreter-independent evaluation tools. Novel web technologies facilitate the design of web interfaces that support online, multiple-choice questionnaires, while exploiting the storage of tracking data as a source of information about user interaction. This paper proposes an online, multiple-choice sign language questionnaire based on an intuitive methodology. It helps users to complete tests and automatically generates accurate, statistical results using the information and data obtained in the process. The proposed system presents SL videos and enables user interaction, fulfilling the requirements that SL interpretation is not able to cover. The questionnaire feeds a remote database with the user answers and powers the automatic creation of data for analytics. Several metrics, including time elapsed, are used to assess the usability of the SL questionnaire, defining the goals of the predictive models. These predictions are based on machine learning models, with the demographic data of the user as features for estimating the usability of the system. This questionnaire reduces costs and time in terms of interpreter dedication, as well as widening the amount of data collected while employing user native language. The validity of this tool was demonstrated in two different use cases.

Original languageAmerican English
Pages (from-to)429-441
Number of pages13
JournalArtificial Intelligence for Engineering Design, Analysis and Manufacturing: AIEDAM
Volume33
Issue number4
DOIs
Publication statusPublished - 2019

Keywords

  • Accessibility
  • HTML5
  • human-computer interaction
  • interface
  • sign language
  • video

Fingerprint Dive into the research topics of 'Design and development of sign language questionnaires based on video and web interfaces'. Together they form a unique fingerprint.

Cite this