Eye and Face Protection Violations Have Ranked Among OSHA’s Top 10 Since 2018: An Evidence-Based Framework for Intelligent Detection, Risk Assessment, and Timely Intervention
International Journal of Scientific Research in Science, Engineering and Technology · 19.11.2026 · 13 (6) · 140-165
DOI: 10.32628/ijsrset2613611 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Science, Engineering and Technology, online erschienen am 19.11.2026. Verfasst von Joseph Olufemi Fasinu. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Eye and face protection violations under the Occupational Safety and Health Administration (OSHA) standard have repeatedly appeared among the most frequently cited federal workplace safety standards since fiscal year 2018, highlighting persistent challenges in personal protective equipment (PPE) compliance, hazard recognition, and occupational safety management. Workers in construction, manufacturing, and other high-risk industrial environments remain vulnerable to eye and facial injuries arising from flying particles, chemical splashes, molten materials, dust, and other hazardous exposures. Although conventional workplace inspections, safety training, and supervisory monitoring remain essential, these approaches are often constrained by limited observation coverage, intermittent assessments, delayed identification of violations, and inconsistent enforcement. Advances in artificial intelligence (AI), computer vision, and edge computing offer opportunities to strengthen PPE monitoring through automated recognition, contextual risk assessment, and timely corrective intervention. This study examines the regulatory, technological, and operational challenges associated with persistent eye and face protection violations and develops an evidence-based framework for integrating AI-assisted PPE detection with proactive occupational safety management. Specifically, it investigates how deep learning-based object detection, worker–PPE association, pose estimation, edge inference, hazard-aware decision rules, and human-centered intervention mechanisms can improve the identification, prioritization, and resolution of potentially unsafe PPE practices. An evidence-based narrative synthesis was conducted using OSHA regulatory information and peer-reviewed literature addressing occupational eye and face injuries, PPE compliance monitoring, computer vision architectures, machine learning, workplace hazard assessment, and safety management systems. Particular attention was given to You Only Look Once (YOLO)-based detection models, pose and key point estimation, lightweight edge-computing architectures, temporal tracking, and multiclass PPE classification. Existing studies were critically examined to identify methodological strengths, operational limitations, performance variability, and research gaps. Building on the reviewed evidence, this study develops a risk-informed occupational eye and face protection assessment methodology that integrates regulatory requirements, hazard exposure, protective equipment adequacy, monitoring reliability, and corrective intervention effectiveness. Rather than proposing another automated PPE detection architecture, the methodology examines how technology-generated observations can be interpreted within task-specific safety requirements and translated into measurable intervention priorities. It distinguishes confirmed equipment absence, incorrect positioning, unsuitable protection, and insufficient visual evidence while incorporating mathematical indicators for hazard-weighted prioritization, detection reliability, corrective-action timeliness, and occupational injury incidence. Computer vision and edge-based monitoring are examined as supporting sources of safety evidence rather than independent injury-prevention solutions. The resulting methodology provides a structured basis for prospective evaluation of monitoring-supported safety interventions. No newly trained detection model, experimental dataset, or completed field trial is reported. Unlike conventional PPE recognition systems that primarily identify equipment presence, the framework emphasizes whether appropriate protective equipment is correctly positioned, associated with the exposed worker, and suitable for the prevailing occupational hazard. It also incorporates human oversight, privacy safeguards, explainable decision support, and mechanisms for managing false-positive detections and ambiguous visual conditions. The study proposes a multidimensional evaluation strategy extending beyond conventional computer vision metrics such as precision, recall, mean average precision, and inference speed. Recommended performance indicators include event-level violation detection sensitivity, false alerts per worker-hour, end-to-end notification latency, time to verified corrective action, unresolved hazardous exposure duration, recurrence of PPE violations, and longitudinal occupational injury outcomes. The proposed framework provides a structured basis for comparing AI-assisted monitoring with conventional safety inspection approaches while distinguishing technical detection performance from operational intervention effectiveness. It further identifies research priorities involving small-object recognition, transparent protective eyewear detection, cross-site generalizability, risk-based alert optimization, worker privacy, and prospective validation of injury-prevention outcomes. AI-assisted eye and face protection monitoring represents a promising approach to addressing persistent PPE compliance challenges through continuous observation, contextual violation assessment, and timely intervention. However, high detection accuracy and rapid alert generation alone do not demonstrate reductions in occupational injuries. Effective prevention requires reliable worker - PPE association, hazard-specific compliance assessment, accountable human intervention, verified corrective action, and integration with established occupational safety procedures. The proposed framework advances the conceptual foundation for transforming automated PPE recognition into a measurable, human-centered workplace safety intervention. Future empirical research should validate the framework across diverse industrial environments and determine whether improved violation detection and corrective-action timeliness translate into sustained reductions in hazardous exposure and preventable eye and face injuries.
Zitieren
APA
Fasinu, J. O. (2026). Eye and Face Protection Violations Have Ranked Among OSHA’s Top 10 Since 2018: An Evidence-Based Framework for Intelligent Detection, Risk Assessment, and Timely Intervention. International Journal of Scientific Research in Science, Engineering and Technology, 13(6), 140-165. https://doi.org/10.32628/ijsrset2613611
Vancouver
Fasinu JO. Eye and Face Protection Violations Have Ranked Among OSHA’s Top 10 Since 2018: An Evidence-Based Framework for Intelligent Detection, Risk Assessment, and Timely Intervention. International Journal of Scientific Research in Science, Engineering and Technology. 2026;13(6):140-165. doi:10.32628/ijsrset2613611
BibTeX
@article{Fasinu2026,
author = {Fasinu, Joseph Olufemi},
title = {Eye and Face Protection Violations Have Ranked Among OSHA’s Top 10 Since 2018: An Evidence-Based Framework for Intelligent Detection, Risk Assessment, and Timely Intervention},
journal = {International Journal of Scientific Research in Science, Engineering and Technology},
year = {2026},
volume = {13},
number = {6},
pages = {140--165},
doi = {10.32628/ijsrset2613611}
}
EUVIMED