Vision Care: A Computational Framework for Distance-Aware Typographic Scaling and Display Adaptation
Shreyash Killerdar · Zenodo (CERN European Organization for Nuclear Research) · 2026
Prolonged work at emissive displays is associated with digital eye strain (DES), a cluster of accommodative, binocular, and ocular-surface symptoms. This paper treats relevant display-user mismatches as a computer-science control problem rather than as a clinical one. It formulates three computable interventions: (i) a pinhole-camera distance estimator driven by a real-time face box, (ii) a closed-loop typographic scale that maintains retinal letter size near the ISO 9241-303 comfort band of 16–22 arcminutes, and (iii) an optional spectral and luminance policy that lowers correlated colour temperature and peak luminance in the evening while preserving contrast. A lightweight prototype couples an OpenCV Haar cascade detector with a banded proportional zoom law and exponential smoothing. Qualitative desk-side trials on commodity laptop hardware were used to assess controller behaviour. No clinical efficacy or quantitative distance-accuracy study was conducted. The contribution is the mapping of established ergonomic constraints onto a real-time vision pipeline.