After clicking “Watch Now” you will be prompted to login or join.


Click “Watch Now” to login or join the NVIDIA Developer Program.


From a Safe City to a Smart City: Thanks to AI and BI Coupled Together on Top of the Urban Video-Protection Infrastructure

Laurent Gomez , SAP SE | Patrick Duverger, Antibes

GTC 2020

Nowadays, every big city is equipped with video-surveillance infrastructure. Police monitor and analyze video streams manually. It's time-consuming and costly, and its efficiency is questionable. Cameras are mostly used to analyze a situation after it occurs. Smart cities want to leverage this infrastructure to improve secure early detection of incidents in public spaces, such as terror attacks or abnormal crowd movement. By empowering cameras with deep-learning capabilities on the edge, we upgrade cameras into multi-function sensors. The city of Antibes (also known for Sophia-Antipolis, the French silicon valley located on the French Riviera) is experimenting with using artificial intelligence to transform urban video-surveillance cameras into internet-of-things sensors that can automatically detect problems and measure the pulse of the agglomeration in real time. This makes it possible to index the city, to switch from video flows meant to be viewed by an operator to textual analytical data that can be stored in databases, in order to make statistical analyses and raise alerts. AI helps us to do the shift from video protection to business intelligence applied to a city. Based on NVIDIA GPUs, the city's cameras can recognize cars and pedestrians to count traffic, or to detect abnormal truck behavior around schools, for example. Being alerted to abnormal situations in real time is a tremendous benefit, but being able to plug all this data into predictive analytics in SAP HANA really transforms a safe city into a smart city.

View More GTC 2020 Content