More and more sophisticated sensors are equipping surveillance systems and generating massive amount of data that need to be analysed and processed, sometimes overwhelming the capabilities of security operators. The development of new techniques based on artificial intelligence (AI), neural networks and deep learning in support of security system, is becoming a key factor for effective solutions. Leonardo has developed innovative platforms, secure-by-design, based on AI for the massive analysis of video and data.

Nowadays it is estimated that more than 1.5 trillion hours of video was captured globally by more than 200 million surveillance cameras. The main parts of these videos are never watched or played back, due to lack of time or resources. As a result, security threats are missed and suspicious behaviours aren’t noticed in time to prevent incidents. AI-based video surveillance allows you to use the system proactively to:

• Analyse large volumes of video and obtain intelligent information by increasing detection, prediction and accuracy
• Receive timely alerts on critical situations and improve security and crime prevention
• Support the task of the control room operators by having more efficient use of resources

Ganimede is the Leonardo platform for the large-scale analysis of live and recorded data streams based on deep learning. Ganimede is implemented exploiting Leonardo’s extensive know-how in video analysis, in IT platforms and security, supported by competence centres specialised in artificial vision and deep learning.

The Ganimede Video Content Analysis platform enhances situational awareness and transforms threat detections from a manual, resource-intensive operation into an efficient and automated process. It is designed and developed to:

• Provide a unique platform for audio/video analysis
• Have a unique framework deployable for data centre, edge computing, automotive