HPCACT-2022: Video Annotation for Anomaly Detection in Data Centers: HPCACT-2022

This groundbreaking code provides a comprehensive annotation and interface to over 10 hours of video activities in high-performance data centers, encompassing over 20 different types of anomalous activities. Enabling machine learning for video surveillance systems in these centers represents a significant breakthrough in the field. The code includes labeled normal activities, such as racking a server and cabling a new system, and anomalous activities, like safety violations, medical emergencies, and equipment failure. Companies maintaining large data centers with video surveillance, including big tech firms and national laboratories, stand to benefit from this technology. With over 10,000 high-performance computing data centers in the US alone, this software has immense potential and should be actively marketed to this industry.

This software is open source and available at no cost.