MoGaze
收藏资源简介:
MoGaze数据集由斯图加特大学创建,专注于日常操作任务中的全身运动数据,包括工作空间几何和眼动追踪。该数据集包含180分钟的运动捕捉数据,记录了1627次拾取和放置动作,使用反射标记和可穿戴瞳孔追踪设备进行数据采集。数据集旨在为设计评估全身运动预测算法提供支持,并展示眼动作为预测人类意图的强大指标。未来计划扩展到包含两人协作任务的数据集。
The MoGaze dataset, created by the University of Stuttgart, focuses on full-body motion data during daily manipulation tasks, including workspace geometry and eye-tracking. It contains 180 minutes of motion capture data, recording 1627 pick-and-place actions collected via reflective markers and wearable pupil-tracking devices. This dataset is intended to support the design and evaluation of full-body motion prediction algorithms, and demonstrates that eye movement serves as a robust indicator for predicting human intent. Future plans include expanding the dataset to cover two-person collaborative tasks.




