VisioPhysio Fatigue Detector Dataset

To support the development and evaluation of VPFD, we construct a multimodal fatigue dataset that integrates both visual and physiological signals in realistic usage scenarios. The dataset is collected using two commercially available wearable devices: Microsoft HoloLens 2 and Google Pixel Watch 2. HoloLens 2 provides rich visual and behavioral information such as gaze direction, blink behavior, eye openness, and head movements, while Google Pixel Watch 2 continuously records physiological responses including heart rate and heart rate variability (HRV).Participants perform sustained attention and cognitively demanding tasks designed to gradually induce different levels of fatigue, during which multimodal data are continuously recorded. Fatigue labels are obtained through task design, performance degradation, and validated subjective self-reports. The resulting dataset captures complementary visual and physiological fatigue patterns and enables comprehensive multimodal fatigue analysis and model training.