Sensors for Human Visual Attention Behaviour Detection and Prediction: A <i>Review</i>

As the primary means of information perception and an indicator of cognitive state, visual attention is a fundamental aspect of human activity. While advancements in manufacturing and artificial intelligence have made sensors capable of detecting and predicting complex visual attention behaviour, a systematic overview of the current state of associated sensor technologies remains notably absent. To address this gap, we adopt the PRISMA-ScR methodology to systematically review 185 peer-reviewed publications from 2018 onwards. Our analysis includes six categories of sensors employed for visual attention behaviour detection and prediction: the mechanical, visual, electrooculography, magnetic, emerging technology, and multi-sensor fusion. We examine five key questions to these sensors’ operation principles, performance features, and representative applications across domains such as driving monitoring, cognitive assessment, virtual reality, and neuroergonomics. While findings reveal trends towards sensor miniaturisation, multimodal integration, and low-power operation, our review confirms challenges in cross-user generalisation, privacy preservation, and long-term stability remain. Consequently, our review indicates a comprehensive picture for future sensor research: pioneering flexible transparent materials, developing self-powered system architectures, and creating adaptive prediction algorithms.