June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
Gender Identification of faces under manipulated ocular optics
Author Affiliations & Notes
  • Clara Benedi-Garcia
    Consejo Superior de Investigaciones Cientificas, Madrid, Spain
    Consejo Superior de Investigaciones Cientificas, Madrid, Spain
  • Miguel P Eckstein
    Vision and Image Understanding Lab, University of California System, Santa Barbara, California, United States
  • Susana Marcos
    Consejo Superior de Investigaciones Cientificas, Madrid, Spain
  • Footnotes
    Commercial Relationships   Clara Benedi-Garcia, None; SARA AISSATI, None; Miguel Eckstein, None; Susana Marcos, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 504. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Clara Benedi-Garcia, SARA AISSATI, Miguel P Eckstein, Susana Marcos; Gender Identification of faces under manipulated ocular optics. Invest. Ophthalmol. Vis. Sci. 2021;62(8):504.

      Download citation file:

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

  • Supplements

Purpose : Correcting ocular aberrations of the eye improves certain visual tasks such as visual acuity and familiar face recognition. In addition, the eye appears to be adapted to its native optics. We evaluated how manipulated aberrations influence face gender identification (GI) performance

Methods : We used a custom-built Adaptive Optics (AO) system, with a Hartmann-Shack wavefront sensor (32x32 microlenses; HASO, ImagineEyes, 827mm), an electromagnetic deformable mirror (DM, 52 actuators; MIRAO), a motorized Badal system and a psychophysical channel (CRT monitor). Nine subjects participated in the study (sph=-1.6±2.15D, cyl=-0.6±0.6D, age=30.4±9.6yrs). The subjects’ ocular aberrations were measured to estimate their Visual Strehl(VS, max 1). Five conditions were evaluated for GI and Visual Acuity (VA): natural aberrations (Nat), AO-correction (AO), 90deg rotated aberrations (Rot), and 2 external patterns, one better (NatB) and worse (NatW). For the GI experiment, we convolved images for each manipulated aberration condition (set of 200male/200female, randomly presented and alternating conditions in blocks of 100 images). Images were presented through the DM correcting subjects’ aberrations. VA was measured with conditions mapped in the DM, using tumbling E letters (8 orientations) and QUEST. Learning effects were analyzed with a logistic regression in a binary model

Results : There was no learning effect through the session. GI performance (%GI) was highly correlated with VS (r=0.90; p<0.005). %GI for the AO vs Nat conditions were highly correlated (r=0.68, p=0.06) with higher differences %GI AO-Nat in subjects with poorer optical quality (≥8%). %GI was higher for Nat than Rot in the less aberrated subjects with a significant %GINat-Rot vs RMS correlation (r=0.4; p=0.33). Only 29% of the subjects with poorer optics (avg VS=0.2±0.06) obtained a higher %GI with Nat than with either NatB (0%) or NatW (67%). However, 86% of the subjects with better optics (avg VS=0.4±0.2) performed better with Nat than with either NatB (100%) or NatW (75%). %GI and VA correlated across conditions at subject level (r=0.42, p<0.02)

Conclusions : Both optical aberrations and neural adaptation to native blur (and its orientation) appear to play a significant role in GI, with a potentially larger role of adaptation to native aberrations in subjects with better optics. VA seems to be a good predictor of GI performance

This is a 2021 ARVO Annual Meeting abstract.


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.