Abstract Unplanned extubation (UEX) in the intensive care unit (ICU) is a serious adverse event. Current prevention strategies relying on staff vigilance have limitations. Artificial intelligence (AI)-based computer vision offers a new approach for real-time, non-contact monitoring. Employing design science, the system was built using an open-source library (MediaPipe). Core modules included human tracking, hand detection, dynamic risk-zone management, and hierarchical alerting.
