SafetyDetect
收藏资源简介:
SafetyDetect数据集是由亚马逊科学和马里兰大学合作创建,旨在帮助家庭机器人识别家中的不安全或不卫生情况。该数据集包含1000个异常家庭场景,每个场景都包含需要检测的不安全或不卫生情况。数据集的创建过程涉及对大量家庭危险统计研究的参考,以及通过用户调查添加的特定案例。SafetyDetect数据集的应用领域主要集中在家庭机器人的异常检测,帮助机器人识别并报告家中的潜在危险,如未关闭的炉灶、易触及的毒物等,从而提高家庭安全性。
The SafetyDetect Dataset was jointly developed by Amazon Science and the University of Maryland, with the goal of assisting home robots in recognizing unsafe and unsanitary conditions within residential environments. This dataset comprises 1000 abnormal household scenarios, each containing unsafe or unsanitary conditions that require detection. During the construction of this dataset, reference was made to a large number of statistical studies on household hazards, and specific cases were added via user surveys. The primary application domain of the SafetyDetect Dataset is anomaly detection for home robots, enabling robots to identify and report potential household hazards such as stoves left on, easily accessible poisons, and the like, thereby improving home safety.

- 1"Don't forget to put the milk back!" Dataset for Enabling Embodied Agents to Detect Anomalous Situations亚马逊科学, 马里兰大学 · 2024年



