Gaze-Regularized VLMs for Ego-Centric Behavior Understanding
This work addresses the challenge of predicting future events with detailed action descriptions for applications in human-computer interaction or surveillance, representing an incremental advance by integrating gaze data into existing VLM frameworks.
The study tackled the problem of enhancing Vision Language Models for egocentric behavior understanding by incorporating eye gaze information, resulting in a nearly 13% improvement in semantic scores compared to baseline models.
Eye gaze, encompassing fixations and saccades, provides critical insights into human intentions and future actions. This study introduces a gaze-regularized framework that enhances Vision Language Models (VLMs) for egocentric behavior understanding. Unlike existing methods that rely solely on visual data and overlook gaze information, our approach directly incorporates gaze information into the VLM architecture during training. By generating gaze-based queries, the model dynamically focuses on gaze-highlighted regions, while a gaze-regularization mechanism ensures the alignment of model attention with human attention patterns. To better understand how gaze can be effectively integrated into VLMs, we conducted extensive experiments exploring various strategies for incorporating gaze data. These innovations enable the prediction of future events with detailed action descriptions. Experimental results demonstrate a nearly 13 % improvement in semantic scores compared to baseline models not leveraging gaze data, highlighting the effectiveness of our approach. This work establishes a foundation for leveraging the human gaze in VLMs, significantly boosting their predictive capabilities in applications requiring accurate and robust future event prediction.