BiCE-HG: An indoor bi-conditional egocentric hand gesture dataset for intelligent reality systems

Robust egocentric hand-gesture recognition remains challenging under the rapidly changing illumination typical of intelligent reality environments. Existing first-person gesture datasets rarely quantify lighting, limiting controlled analysis of illumination robustness. Unlike prior egocentric datasets that describe lighting qualitatively (e.g., “indoor”, “bright”), BiCE-HG is the first to provide calibrated lux measurements at 80 spatial points, enabling reproducible illumination robustness benchmarking. BiCE-HG (Bi-Conditional Egocentric Hand Gestures) dataset addresses this gap as the first egocentric gesture dataset with measured bi-conditional lighting and spatial illuminance documentation. BiCE-HG contains 19 gestures performed by 23 participants across six conditions combining sitting, standing, and walking with full and low lighting setups. Walking trials capture realistic motion through a 4 m lighting corridor with transitions between bright and dim zones. Spatial illuminance is recorded every 25 cm using a calibrated meter, enabling precise modelling of lighting gradients and reproducible experimental conditions. Baseline evaluation reaches 81.24% validation accuracy (F1 = 0.80) on the static (single-frame) release with a no-attention ST-GCN, the dynamic (temporal) release lifts accuracy to 88.57% with the identical architecture, isolating a 7.3-percentage-point gain from temporal information alone, while adding tri-attention contributes a further marginal gain to 89.24% (F1 = 0.88). Lighting-robustness analysis shows stable spatial measurements, with a coefficient of variation below 5% for most conditions across a 4.95x illuminance range (29–523 lux). BiCE-HG provides quantified environmental metadata and realistic mobility lighting interactions, supporting the development of illumination-robust egocentric gesture-recognition systems for AR/VR applications.The dataset is publicly available at https://doi.org/10.15131/shef.data.32374137.