Scene Text Access: A Comparison of Mobile OCR Modalities for Blind Users

Area of Science:

• Assistive Technology
• Human-Computer Interaction
• Computer Vision

Background:

• Mobile Optical Character Recognition (OCR) apps offer potential for visually impaired individuals to access printed text.
• Existing OCR applications vary in their user interface and text acquisition strategies, impacting usability for blind users.

Purpose of the Study:

• To evaluate and compare the performance and usability of three distinct mobile OCR applications for blind participants in indoor settings.
• To identify effective user strategies and inherent challenges associated with using mobile OCR technology without visual feedback.

Main Methods:

• A study involving seven blind participants using three mobile OCR applications: Microsoft SeeingAI (Short Text mode), Spot+OCR, and Guided OCR.
• Quantitative data collection included true positive ratios and traversal speed.
• Qualitative data gathered through observations and an exit survey to assess user experience and challenges.

Main Results:

• Performance metrics (true positive ratio, traversal speed) were recorded for each application.
• User strategies for text detection and acquisition varied across participants and apps.
• Qualitative feedback revealed specific usability challenges and preferences related to interface design and feedback mechanisms.

Conclusions:

• Different mobile OCR app designs present unique advantages and disadvantages for blind users.
• Understanding user strategies and challenges is crucial for developing more effective and accessible OCR solutions for the visually impaired.
• Further research is needed to optimize mobile OCR technology for seamless indoor navigation and text access.