Eye Image-based Algorithms to Estimate Percentage Closure of Eye and Saccadic Ratio for Alertness Detection
This work addresses the problem of detecting operator fatigue for safety-critical applications, representing an incremental improvement with specific technical innovations.
The paper developed two image-based algorithms to estimate Percentage Closure of Eyes (PERCLOS) and Saccadic Ratio (SR) for detecting alertness levels, achieving higher accuracy than existing methods through a novel combination of gray scale and Near Infrared camera with passive NIR illuminator.
The current research work has developed two novel algorithms for image-based measurement of Percentage Closure of Eyes-PERCLOS and Saccadic Ratio-SR. The PERCLOS is estimated by correlation filter-based technique. An innovative combination of gray scale and Near Infrared sensitive camera with passive NIR illuminator helps to achieve higher accuracy than the existing art. Two novel techniques have been developed for the detection of iris centre and eye corners. We propose an index called Form Factor to find the iris position. The saccadic velocity profile can be estimated from the temporal information of the iris positions using standard tracking algorithm such as Extended Kalman filter. Experimental results indicate that the estimation of both SR and PERCLOS can predict the level of alertness of an operator from onset of diminished alertness to fatigue.