Real-Time Iris Tracking-based on a Generalized Probabilistic Particle Filter

Objectives: We have proposed a new approach for tracking the iris, based on a generalized probabilistic particle filter. Accurate iris tracking plays an important role in many advanced applications, such as human-computer interactions and driver fatigue detection. Methods/Statistical Analysis: This paper illustrates a new method for iris tracking, based on a generalized probabilistic particle filter. This approach utilizes a sample set for the tracked iris, which is constructed at the beginning of the tracking process. The prior representation and position of the tracked iris were then predicted, based on minimization of the parameters of the proposed generalized probabilistic distribution. Findings: The computation of the likelihood of generalized particle filters for each distribution function is hen estimated. Our approach had provided reliable results with a low error rate and low levels of computation. This technique aimed to improve the speed of the eye tracking algorithm so that can be used for real-time applications. Application/Improvements: Accurate iris tracking plays an important role in many advanced applications, such as human-computer interactions and driver fatigue detection.