Eye Open/Close Detection and Eye State Recognition Using Circularity Metric

In this paper, we present a MATLAB-based approach for detecting open and closed eyes by leveraging circularity metrics to identify eye states. The implementation begins with eye image acquisition, followed by comprehensive image processing algorithms for analysis. Key MATLAB functions employed include imread() for image loading, rgb2gray() for grayscale conversion, and imbinarize() for threshold-based segmentation. The core algorithm calculates circularity using the formula: Circularity = (4π × Area) / (Perimeter²), where regionprops() function extracts geometric properties from binary eye regions. Higher circularity values indicate open eyes (more circular iris visibility), while lower values suggest closed states. This method proves valuable in various applications such as driver fatigue monitoring systems, sleep studies, and human-computer interaction interfaces. By implementing morphological operations like imopen() and imclose() for noise reduction, the system achieves robust state classification, enhancing both safety and operational efficiency in real-world scenarios.