International Journal of Environmental Research and Public Health · 5 authors, 1 centre
This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.
This exploratory study found that average blinking rate, measured via a computer webcam in real-life settings, is a significant predictor of computer vision syndrome (CVS). For each additional blink per minute, the CVS score decreased by 1.26 points. These findings support the development of real-time blink detection systems to monitor and potentially reduce CVS risk during computer use.
**Background:** Computer vision syndrome (CVS) is a group of eye and vision-related problems resulting from prolonged use of digital screens, affecting 60-70% of office workers. A key risk factor is reduced blink rate during computer work, which can lead to dry eyes and ocular discomfort. While existing tools assess CVS via questionnaires, there is a need for unobtrusive, real-time detection systems that can monitor risk factors during daily computer use. This study aimed to investigate whether blinking data collected from a computer webcam in real-life settings can serve as a reliable indicator for predicting CVS.
**Methods:** This longitudinal exploratory study included 13 students (final sample: 11 after exclusions) aged 21-25 years (mean 22.0 ± 1.0), with 9 females and 2 males. Participants used the 'HealthyRoad' webcam-based software during scientific writing tasks for at least 4 hours daily. The software collected blink data without storing images or videos, outputting numerical data. Blink rate was computed from the number of blinks, frame rate (4-30 fps), and acquisition rate. CVS was assessed using the validated Computer Vision Syndrome Questionnaire (CVS-Q), which evaluates frequency and intensity of 16 symptoms; a score ≥6 classified a participant as a CVS case. Data files with <2 minutes or <500 KB were excluded, and only files with ≥1 hour of screen time and ≥75% face detection were retained. Statistical analyses included descriptive statistics, independent t-tests, Pearson correlations, and a regression model with bootstrapping.
**Key Results:** The mean CVS score was 18.09 ± 6.09. CVS was negatively and highly correlated with average blinking rate (r = -0.781, p = 0.005, 95% CI [-0.927, -0.248]) and with age (r = -0.639, p = 0.034, 95% CI [-0.957, -0.148]). Gender was not significantly associated with CVS (p = 0.342). The regression model (F(3,7) = 7.685, p = 0.013, R² = 0.876) showed that average blinking rate was the only significant predictor of CVS (B = -1.262, p = 0.019, 95% CI [-2.249, -0.275]), meaning each additional blink per minute was associated with a 1.26-point decrease in CVS score. Age and gender were not significant predictors.
**Clinical Implications:** This study provides preliminary evidence that real-time blink detection via a standard webcam can predict CVS severity. The strong inverse relationship between blink rate and CVS score suggests that monitoring blink rate could enable early detection and intervention. Such a system could offer real-time feedback (e.g., reminders to blink or take breaks) to reduce CVS symptoms and improve ocular comfort during computer work. However, the small sample size and gender imbalance limit generalizability. Future research with larger, more diverse samples and additional eye variables (e.g., gaze, pupil diameter) is needed to validate and refine the algorithm for broader occupational health applications.