User identification for home entertainment based on free-air hand motion signatures

Omri Mendels, Helman Stern, Sigal Berman

Research output: Contribution to journalArticlepeer-review

18 Scopus citations

Abstract

A user identification system based on free-air hand signature-gestures acquired with a 3-D camera was developed. In the system, users interactively defined their own motion signatures by demonstration and trained the system by performing a plurality of signatures. The system identifies the user by comparing the distances of a sample to other signatures. The distance metric is learned by using neighborhood components analysis. An interactive enrollment algorithm which uses sequential clustering, and allows the system to advise the user during signature selection and system training was developed. Four validation tests were conducted: 1) all users using a single predefined signature-gesture (independent); 2) each user using a personal signature-gesture (dependent); 3) copycat tests for examining robustness against forgery; and 4) operation of the interactive enrollment system. For identifying a single user out of user cohorts of three to seven people, the independent system had average accuracies of 91%-77% depending on cohort size and signature shape. Higher average accuracies of 98%-92% were obtained for the dependent system. In the forgery tests, users with high signature variability over time were susceptible to forgeries, but users with low signature variability obtained a low equal error rate of 0.083. The interactive enrollment system significantly improved recognition accuracy. The proposed system can be integrated into gesture-based home entertainment systems and used for interface customization, content adaptation, and parental control. User attitudes toward the system within this context were assessed using the widely accepted technology acceptance model. Based on 69 responders, the results indicated a positive user attitude toward the system and a high intention to use it. The users expressed a preference for personalized gestures, a finding that indicates the importance of the interactive enrollment module for personalizing the signature-gestures.

Original languageEnglish
Article number6842691
Pages (from-to)1461-1473
Number of pages13
JournalIEEE Transactions on Systems, Man, and Cybernetics: Systems
Volume44
Issue number11
DOIs
StatePublished - 1 Nov 2014

Keywords

  • Clustering methods
  • free-air hand signature gestures
  • human-machine interaction
  • motion analysis
  • user identification

ASJC Scopus subject areas

  • Software
  • Control and Systems Engineering
  • Human-Computer Interaction
  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'User identification for home entertainment based on free-air hand motion signatures'. Together they form a unique fingerprint.

Cite this