Eye Movement Recognition and Gaze Point Prediction with Webcam

With the advancement of human-computer interaction, virtual reality, and related fields, gaze prediction, as a core technology, has been widely applied in various scenarios. Traditional gaze prediction methods mostly rely on expensive hardware devices such as infrared cameras, which limits their practical applications. To address this issue, this paper proposes a low-cost gaze prediction method based on a regular camera. By integrating multi-modal feature fusion of head pose and eye features, along with a lightweight multi-layer perceptron (MLP) model, we achieve a balance between high accuracy and real-time performance without relying on additional hardware. Experimental results demonstrate that the proposed method can achieve a gaze prediction accuracy of 3.95° using a regular laptop camera and only contains 7.73 million parameters.These experiments verify the effectiveness and feasibility of the proposed method on low-cost devices, providing a new solution for gaze prediction in practical applications.