September 2015
Volume 15, Issue 12
Free
Vision Sciences Society Annual Meeting Abstract  |   September 2015
Caricaturing improves face identity recognition in simulated prosthetic vision
Author Affiliations
  • Jessica Irons
    The Australian National University ARC Centre for Cognition and Its Disorders
  • Tamara Gradden
    The Australian National University
  • GuanLing Zhang
    The Australian National University
  • Xuming He
    National Information and Communication Technology Australia (NICTA) Bionic Vision Australia
  • Nick Barnes
    National Information and Communication Technology Australia (NICTA) Bionic Vision Australia
  • Elinor McKone
    The Australian National University ARC Centre for Cognition and Its Disorders
Journal of Vision September 2015, Vol.15, 1208. doi:https://doi.org/10.1167/15.12.1208
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jessica Irons, Tamara Gradden, GuanLing Zhang, Xuming He, Nick Barnes, Elinor McKone; Caricaturing improves face identity recognition in simulated prosthetic vision. Journal of Vision 2015;15(12):1208. https://doi.org/10.1167/15.12.1208.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

The “bionic eye” is a prosthetic device that can restore limited vision to those with degenerative blindness. By directly stimulating undamaged neurons in the visual pathway, the bionic eye allows the wearer to perceive an array of discrete spots of light ("phosphenes"). The resolution of current devices is limited, and performance on complex visual tasks such as face recognition is poor. Thus, researchers have sought image manipulation techniques to enhance visual processing. We explored a technique called caricaturing, aimed at improving face recognition by enhancing high-level face shape information. Caricaturing involves morphing an individual face away from an average face (matched on age, sex, race), exaggerating identity-specific aspects of the individual’s face shape. We tested caricaturing in simulated prosthetic vision, by converting face photographs to arrays of artificial phosphenes (“phosphenising”) at four resolutions: 16x16, 32x32 and 40x40 all with 30% dropout (simulating electrode failure), and 40x40 without dropout. Caricaturing improved both the perceptual individuation and recognition of phoshenised faces. In the perceptual individuation task, participants were shown pairs of faces and rated how different the two individuals appeared. Caricaturing increased the perceived differences between individuals at all resolutions except 16x16. To assess face recognition, participants learned faces in normal high resolution and performed an old/new recognition task on phosphenised faces. Caricaturing improved recognition accuracy for faces at 40x40 without dropout. Recognition at lower resolutions was at chance. We then modified the task to more accurately simulate the experiences of bionic eye patients, by enabling the phosphene array to scan across the face stimuli, and presenting both learning and test faces at the same resolution. Recognition at 40x40 with 30% dropout improved to above chance and now showed a caricature benefit in recognition accuracy, confidence, and scanning time. We conclude that caricaturing offers practical benefits for face recognition with the bionic eye.

Meeting abstract presented at VSS 2015

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×