Determining Eye Blink Rate Level Utilizing Sitting Postural Behavior Data

Abstract

Dry eye syndrome (DES) affects many white-collars workers worldwide. Though it is known that low eye blink rate (EBR) is associated with the risk of DES, it is difficult to improve EBR through self-correction. One way to increase EBR is to warn the worker of low EBR using an external system. Existing EBR measurement devices have limitations, such as physical discomfort or invasiveness, which hinder their acceptance. For a solution that overcomes these limitations, this study aimed to develop a classification system that differentiates the levels of EBR using posture and postural variability data obtained from chair-embedded distance and pressure sensors. Additionally, this study attempted to investigate the relationship between EBR, posture, and postural variability. Participants completed three seated computer tasks, in which eye blink and postural sensor data were collected. The EBR classification system was developed by using a machine learning method

the accuracy of the EBR classification system was 93% across the three task types and study participants. The low EBR level was found to be associated with smaller postural variability and a tendency for the worker to hold a forward-leaning sitting posture. The EBR classification system developed in this study is expected to contribute to the prevention of DES.