This paper introduces a robust approach for eye detection in images, combining traditional techniques and deep learning. Initial preprocessing enhances image quality, followed by Haar- like features and AdaBoost for candidate region detection, optimizing computational efficiency. Subsequent refinement employs a convolutional neural network trained on annotated data for precise eye localization, ensuring resilience to various image conditions. Experimental validation on benchmark datasets demonstrates competitive performance and real-time suitability for applications such as gaze tracking and driver assistance. The proposed method offers a balanced trade-off between accuracy, efficiency, and robustness, making it applicable to diverse computer vision tasks requiring reliable eye detection.

#