SituFont: A Just-in-Time Adaptive Intervention Interface for Enhancing Mobile Readability in Situational Visual Impairments

Paper Title

SituFont: A Just-in-Time Adaptive Intervention Interface for Enhancing Mobile Readability in Situational Visual Impairments

Publication Info

• Topic area: Adaptive typography for mobile readability under situational visual impairments.
• Keywords: Situational visual impairments, adaptive typography, mobile readability, just-in-time interventions, human-in-the-loop, context-aware systems, font adaptation, accessibility, smartphone sensors, user experience.

Background and Problem

• Problem / challenge: Existing solutions for situational visual impairments (SVIs) rely on static or manual adjustments, which are insufficient for dynamic contexts involving factors like motion, lighting, and fatigue. These approaches impose cognitive and interactional burdens on users.
• Significance: Addressing SVIs is crucial for improving mobile readability, reducing cognitive load, and enhancing accessibility in everyday dynamic environments.
• Motivation and related work: Prior work has explored static font adjustments, brightness modeling, and auditory substitutes, but these solutions fail to adapt dynamically to fluctuating conditions or account for individual user preferences. This paper builds on research in adaptive typography and JITAI systems to address these gaps.

Solution

• Proposed approach: SituFont, a context-aware, human-in-the-loop adaptive typography system that dynamically adjusts font parameters (e.g., size, weight, spacing) based on real-time environmental conditions and user feedback.
• Novelty:
  1. Integration of smartphone sensors and human-in-the-loop feedback for personalized, just-in-time font adaptation.
  2. Empirical insights into how environmental, personal, and informational factors affect mobile readability under SVIs.
  3. A hierarchical label tree structure for efficient context representation and adaptation.
  4. Design considerations for reducing cognitive load and enhancing user experience in dynamic reading contexts.
• Procedure and key techniques:
  1. Conducted formative interviews (N=15) to identify key SVI factors and coping strategies.
  2. Performed controlled experiments (N=18) to quantify the impact of environmental factors on font parameter preferences.
  3. Designed and implemented SituFont with components including a machine learning model, a label tree for context representation, and a human-AI loop for feedback-driven personalization.
  4. Evaluated SituFont through a comparative user study (N=12) across eight simulated SVI scenarios.

Results

• Concrete findings:
  • SituFont improved reading goodput (characters per minute) across most SVI scenarios, with significant gains in multi-factor conditions (e.g., intense brightness + high vibration + distraction).
  • Reading comprehension accuracy remained stable across both SituFont and the baseline system.
  • SituFont reduced perceived mental and physical workload, particularly under conditions involving brightness, vibration, and fatigue.
  • User experience ratings showed higher efficiency, supportiveness, and novelty for SituFont compared to the baseline.
• Advantage over baselines:
  • SituFont outperformed a manually adjustable baseline in most scenarios, particularly in dynamic and multi-factor conditions.
  • Dynamic adjustments provided benefits beyond what static optimization could achieve, especially in maintaining reading flow under fluctuating conditions.
• Experiments / evaluation:
  • Conducted a within-subject user study with 12 participants under eight SVI scenarios, including tasks for reading aloud and comprehension.
  • Metrics included reading goodput, comprehension accuracy, NASA-TLX workload scores, and user experience questionnaires (UEQ-S, SUS).
• Limitations and future work:
  • Limited generalizability to older adults, non-Chinese scripts, and long-form reading tasks.
  • Cold-start problem during initial use due to insufficient user data.
  • Privacy, energy, and resource costs not systematically evaluated.
  • Future work should explore richer feedback mechanisms, longitudinal studies, and privacy-preserving deployments.

Summary

SituFont is a context-aware, human-in-the-loop adaptive typography system designed to enhance mobile readability under situational visual impairments. By leveraging smartphone sensors and user feedback, it dynamically adjusts font parameters to accommodate changing environmental and personal conditions. Empirical studies demonstrated significant improvements in reading efficiency and user experience, particularly in complex SVI scenarios. While limitations include cold-start challenges and resource considerations, SituFont provides a robust framework for adaptive typography, with potential applications in broader accessibility and dynamic reading contexts.