A Multimodal EEG-Camera Dataset for Imagined Driving Action Recognition

<p dir="ltr">This dataset contains synchronized EEG and camera-derived visual features collected for imagined driving action recognition.</p><p><br></p><p dir="ltr">The dataset includes recordings from 20 participants performing five imagined driving conditions: left_down, left_up, right_down, right_up, and static.</p><p><br></p><p dir="ltr">The released dataset contains 4,500 paired EEG-camera trials. Each trial includes a 32-channel EEG segment and corresponding camera-derived visual features.</p><p><br></p><p dir="ltr">For privacy protection, raw facial videos are not released.

Instead, de-identified camera-derived visual features are provided.</p><p><br></p><p dir="ltr">Dataset organization:</p><p><br></p><p dir="ltr">EEG/</p><p dir="ltr">Subject-wise EEG recordings</p><p><br></p><p dir="ltr">Camera/</p><p dir="ltr">Subject-wise camera feature sequences</p><p><br></p><p dir="ltr">labels.csv</p><p dir="ltr">Trial-level labels and modality correspondence information</p><p><br></p><p dir="ltr">The dataset is released to facilitate reproducible research on multimodal EEG-camera fusion and imagined driving action recognition.</p>