Video analytics gets boost from AI

Karuppiah talks about NVIDIA's IVA offerings.

Intelligent video analytics (IVA) powered by artificial intelligence (AI) is taking off.

At NVIDIA's AI Innovation Day, Ettikan Karuppiah, Director, Developers Ecosystem, NVIDIA, shared that the AI city opportunity alone is worth US$2 trillion, involving some 1 billion sensors and cameras as well as 250 million hours spent reviewing videos every year. Unlike humans, AI systems can work 24x7 and analyse more videos. Such superior analytics processing can eliminate bottlenecks in access control, understand traffic congestion, and enhance quality control in manufacturing, he noted.

"IVA has impacted many industries, and different types of use cases," he said. Use cases that Karuppiah shared included automatically catching cars which are driving in the wrong direction; testing promotions and new products; or identifying corrosion in infrastructure. AI can help to identify more incidents and insights, more accurately.

NVIDIA has created end-to-end platform support for AI, he shared. For IVA opportunities it has Metropolis, which works together from edge to cloud with NVIDIA's EGX edge computing hardware and third-party solutions to create smarter cities.

To start an AI project, Karuppiah recommends:

- Crafting a clear AI strategy. "You need to know the key projects that you want to start with," he said.

"You must have the right AI software and hardware and the partners to help you."

- Sourcing independent software vendors/systems integration (ISV/SI) partners to engage and work with you, as you cannot do it by yourself.

- Identifying key metrics/key performance indicators which would make operations a success. Be honest about proof of value - the value actually gained by solving a particular problem. Hold a workshop on use cases with NVIDIA or a partner and prioritise them.

- Build scalability in from day one.