A human video database for facial feature detection under spectacles with varying alertness levels

Pressing demand of workload along with social media interaction leads to diminished alertness during work hours.Researchers attempted to measure alertness level from various cues like EEG, EOG, Video-based eye movement analysis,etc. Among these, video-based eyelid and iris motion tracking gained much attention in recent years. However, most of theseimplementations are tested on video data of subjects without spectacles. These videos do not pose a challenge for eye detectionand tracking. In this work, we have designed an experiment to yield a video database of 58 human subjects wearing spectaclesand are at different levels of alertness. Along with spectacles, we introduced variation in session, recording frame rate (fps),illumination, and time of the experiment. We carried out analysis to detect the reliableness of facial and ocular features likeyawning and eyeblinks in the context of alertness level detection capability. Also, we observe the influence of spectacles on ocularfeature detection performance under spectacles and propose a simple preprocessing step to alleviate the specular reflectionproblem. Extensive experiments on real-world images demonstrate that our approach achieves desirable reflection suppressionresults within minimum execution time compared to the state of the art.

日益增长的工作负荷与社交媒体交互需求，导致职场人员在工作时段的警觉性显著降低。研究人员尝试通过多种线索开展警觉性水平测量，例如脑电图（EEG）、眼电图（EOG）、基于视频的眼动分析等。其中，基于视频的眼睑与虹膜运动追踪技术近年来受到广泛关注。然而，此类检测方案大多基于未佩戴眼镜的受试者视频数据进行测试，此类视频不会对眼部检测与追踪构成挑战。本研究设计了一项实验，构建了包含58名佩戴眼镜且警觉性水平各异的受试者的视频数据库。除受试者佩戴眼镜外，我们还在实验中引入了实验会话、录制帧率（fps）、光照条件以及实验时段的变量变化。我们开展了分析，以验证打哈欠、眨眼等面部与眼部特征在警觉性水平检测中的可靠性。此外，我们还分析了佩戴眼镜对眼部特征检测性能的影响，并提出了一种简单的预处理步骤，以缓解镜面反射问题。在真实场景图像上开展的大量实验表明，相较于当前技术前沿方案，本方法能够在最短执行时间内实现理想的反射抑制效果。