Related Experiment Video
Updated: Nov 5, 2025
05:40
Using an Automated Hirschberg Test App to Evaluate Ocular Alignment
Published on: March 24, 2020
14.8K
What Visual Targets Are Viewed by Users With a Handheld Mobile Magnifier App
1Schepens Eye Research Institute, Massachusetts Eye and Ear, Department of Ophthalmology, Harvard Medical School, Boston, MA, USA.
Translational Vision Science & Technology
|May 18, 2021
Summary
Visually impaired individuals using mobile magnifier apps frequently view nontextual objects like scenes and people. This highlights a key need for assistive technology development to support diverse visual requirements.
Area of Science:
- Computer Vision
- Human-Computer Interaction
- Assistive Technology
Background:
- Mobile video magnifier applications are crucial tools for visually impaired individuals.
- Understanding user visual targets informs low-vision research and assistive technology.
- This study analyzes daily activities captured by magnifier app users.
Purpose of the Study:
- To identify and categorize the common visual targets of mobile video magnifier app users.
- To inform the development of more effective assistive technologies for the visually impaired.
Main Methods:
- Utilized an iOS magnifier app to capture 152,819 images from 24,295 users over one month.
- Employed Azure Computer Vision for object recognition and Umeng analytics for data aggregation.
- Analyzed object tags, categorizing over 1300 types into 11 groups, comparing users with and without vision-accessibility features.
Main Results:
- Nontextual objects (indoor scenes, humans, art) constituted 60-70% of visual targets for both accessibility and nonaccessibility user groups.
- Over 40% of images contained multiple object categories.
- While accessibility users viewed text more often, category rankings were similar across groups.
Conclusions:
- Nontextual objects represent a significant portion of the visual needs for individuals using magnifiers.
- Future low-vision research and assistive technology should prioritize support for nontextual object recognition and viewing.
