Efficient Real-Time Eye Gaze Tracking Detection for Human-Computer Integration Using Advanced Techniques

The eye-gaze tracking for detecting different types of eye movements in a continuous stream of gaze data is limited, as it involves Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) features. These features employ individual detectors, each capable of detecting a single movement. However, in some applications, eye tracking may not occur due to issues such as overfitting and noise in image detection. These challenges are addressed in our method, which utilizes CNN and RNN. The goal is to design algorithms that can accurately differentiate between nuanced emotional states, thereby enhancing the precision of eye tracking. To achieve this objective, we employ the CNN State, which is trained on a dataset of human eye images captured by intelligent eyeglasses to obtain an eye state recognition model. This allows for efficient transfer learning, enabling the eye emotion tracking model to benefit from knowledge gained in other domains, even with limited labeled emotion data. It's worth noting that RNNs have a limited memory capacity, which can hinder their ability to effectively capture and retain information over extended periods. Nevertheless, our experimental results demonstrate high-level performance on the FER2013 dataset, achieving an accuracy of 0.97. This performance surpasses other existing models such as ResNet, K-Nearest Neighbors, and VGG19.