Investigative Ophthalmology & Visual Science Cover Image for Volume 61, Issue 7
June 2020
Volume 61, Issue 7
Free
ARVO Annual Meeting Abstract  |   June 2020
Understanding the use of artificial intelligence based visual aids for people with visual impairments
Author Affiliations & Notes
  • Elizabeth Kupferstein
    Information Science, Jacobs Technion-Cornell Institute, Cornell Tech, Cornell University, New York, New York, United States
  • Yuhang Zhao
    Information Science, Cornell Tech, New York, New York, United States
  • Shiri Azenkot
    Information Science, Jacobs Technion-Cornell Institute, Cornell Tech, Cornell University, New York, New York, United States
  • Hathaitorn Rojnirun
    Information Science, Cornell Tech, New York, New York, United States
  • Footnotes
    Commercial Relationships   Elizabeth Kupferstein, None; Yuhang Zhao, None; Shiri Azenkot, None; Hathaitorn Rojnirun, None
  • Footnotes
    Support  none
Investigative Ophthalmology & Visual Science June 2020, Vol.61, 932. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Elizabeth Kupferstein, Yuhang Zhao, Shiri Azenkot, Hathaitorn Rojnirun; Understanding the use of artificial intelligence based visual aids for people with visual impairments. Invest. Ophthalmol. Vis. Sci. 2020;61(7):932.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose :
The purpose of this study is to assess how people with visual impairments use an artificial intelligence (AI) based visual aid. This study will identify which factors to focus on to refine the helpfulness and enhance the user experience of AI-based visual aids.

Methods : We are conducting a diary study with users of Seeing AI, a smartphone application that uses computer vision to recognize visual information in the environment and verbally relate it to a visually impaired user. The diary study lasts 2 weeks, where participants use the Seeing AI application on a daily basis and complete an online questionnaire for each photo taken with the application. The questionnaire asks about the user’s perceived accuracy, content completeness, and descriptiveness of the AI recognitions. We require each participant to use Seeing AI and complete the questionnaires at least 20 times during the study. We end the study with a closing interview allowing the user to elaborate on their questionnaire responses while further exploring their perception of Artificial Intelligence.

Results : We have thus far collected 87 questionnaires from the first 4 participants. All participants have functional vision that is hand motion or less. We identified two primary use cases for Seeing AI: when users seek information they want to keep private from peers and objects they are unable to physically touch (i.e., objects behind a glass). The responses to 5-point Likert scales in the questionnaires were 58.6% positive for the content completeness of the AI recognitions and 47.1% positive for the descriptiveness. Nevertheless, satisfaction with the recognition results were only 43.0% positive. Participants elaborated that the lack of visual details meant the results were not more useful than gathering information from touching the object (i.e., the application recognized the TV but did not recognize the images on the TV). Moreover, they explained that they trusted the recognition responses more if the response contained more details.

Conclusions : This preliminary study shows that people with visual impairments want to use AI-based visual aids to gain visual information that is unattainable via tactile methods. The descriptiveness of the AI recognitions was identified as a factor that both increases the usefulness and trustworthiness of AI, assuaging the doubt caused by the potential for computer errors.

This is a 2020 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.

×