Event-based & High-FPS Hand Gestures

This dataset contains synchronized recordings of eight distinct hand gestures captured simultaneously with a high-speed camera (1000 FPS) and an event-based camera. The recorded gestures include: anticlockwise turn, clockwise turn, full hand pinch, fingers pinch, left slide, right slide, upward slide, and downward slide. Each gesture was performed in multiple sessions, with around 80 paired recordings per gesture, resulting in a comprehensive collection suitable for developing and benchmarking multi-modal deep learning models for gesture recognition. The dataset also includes dedicated recordings for position and time calibration between the two camera modalities. Detailed documentation is provided in dataset_description.docx, and the dataset itself is available in recordings.zip, which can be reassembled and extracted using: cat recordings.tar.gz.* | tar -xvzf -本数据集包含8种不同手部手势的同步录制数据，由高速相机（1000 FPS）与基于事件的相机（event-based camera）同步采集得到。所录制的手势包括： 逆时针旋转、 顺时针旋转、 全手捏合、 指尖捏合、 左滑、 右滑、 上滑以及下滑。 每种手势均开展了多轮录制，每类手势约包含80组配对录制样本，由此构建出一套适用于手势识别多模态深度学习模型开发与基准测试的完备数据集。此外，本数据集还涵盖了用于两种相机模态间位置与时间校准的专用录制数据。 详细说明文档已收录于dataset_description.docx，数据集本体存储于recordings.zip，可通过如下命令完成拼接与解压：cat recordings.tar.gz.* | tar -xvzf -