3ET: Efficient Event-based Eye Tracking using a Change-Based ConvLSTM Network

Abstract--This paper presents a sparse Change-Based Convolutional Long Short-Term Memory (CB-ConvLSTM) model for event-based eye tracking, key for next-generation wearable healthcare technology such as AR/VR headsets. Utilizing a delta-encoded recurrent path enhancing activation sparsity, CB-ConvLSTM reduces arithmetic operations by approximately 4.7 without losing accuracy when HE process of eye movements often reveals our mental processes and comprehension of the visual realm. Implementing eye tracking technology offers many possibilities in Eye tracking is a significant field in computer vision [8]- augmented reality/virtual reality (AR/VR) domains, enabling [10], yet it's relatively unexplored with event cameras due to techniques like foveated rendering to offer a more compelling the scarcity of relevant event-based datasets [11], [12]. Eye tracking has common approaches guide recent advances in event-based eye potential benefits in wearable healthcare applications. For tracking algorithms, mirroring those of traditional computer instance, it can aid in identifying eye movement disorders associated vision: (1) The 3D model-based method locates key points with diseases like Parkinson's or Alzheimer's, thereby corresponding to the image's geometrical features and fits enabling early diagnosis and regular assessments [3], [4].