Query-driven video event processing for the internet of multimedia things

Advances in Deep Neural Network (DNN) techniques have revolutionized video analytics and unlocked the potential for querying mining event patterns. This paper details GNOSIS, an processing platform to perform near-real-time detection a distributed setting. GNOSIS follows serverless approach where its component acts as independent microservices can be deployed at multiple nodes. uses declarative query-driven users write customize queries spatiotemporal reasoning. The system converts incoming streams into continuous evolving graph stream using machine learning (ML) DNN models pipeline applies matching pattern detection. both stateful stateless matching. To improve Quality of Service (QoS), recent work incorporates optimization like adaptive scheduling, energy efficiency, content-driven windows. demonstrates Occupational Health Safety query use cases show efficacy.