VR scene optimization algorithm driven by eye tracking data

In virtual reality (VR) scene optimization based on eye tracking data, how to accurately capture the user’s eye movement data and optimize the scene presentation accordingly is one of the key challenges. In this paper, a new approach combining Pupil centered Cornea Reflection (PCCR) with Deep Q-Networks (DQN), named pupil-Cornea Q-Networks (PCQN), is proposed. PCQN uses PCCR technology to accurately detect the user’s eye movement trajectory, and uses DQN to learn the optimal scene rendering strategy. Specifically, PCCR provides high-precision eye tracking data, while DQN dynamically adjusts the rendering parameters of VR scenes based on this data for a smoother, more realistic user experience. Finally, this paper compares the model architecture with the traditional optimization algorithms, and the results show that PCQN has significant advantages in improving rendering efficiency, robustness and user satisfaction. This model is proved to be feasible and effective in VR scenes driven by eye tracking data.