Xjera Labs zooms in to AI to enhance security

Chu: "NVIDIA GPUs compared to CPUs are much more cost effective."
Chu: “NVIDIA GPUs compared to CPUs are much more cost effective.”

Security is a growing concern among governments and organisations of all sizes. They must balance the need to provide access to the right people while keeping suspicious folks at bay. Any lapse can result in dire consequences that impact confidence in the country or company.

To tackle this challenge, Xjera Labs has turned to artificial intelligence (AI) powered by the graphics processing unit (GPU). It has developed AI-based image and video analytics (VA) solutions for commercial applications.

Established in 2013, the Singapore-based company provides VA services to a local theme park, a chain of sushi restaurants, logistics companies, shopping malls, and government agencies.

“Currently, most of our projects are in public safety especially with governments, like in Singapore. Other projects are with multinational corporations in real estate and security,” said Ethan Chu, CTO and Founder of Xjera Labs.

“Our company impacts industries mainly by reducing labour costs for our partners. Previously, our partners must hire a lot of human resource to do things such as accounting and looking at computer screens to look for problems. We use AI for computer vision so that our partners do not need to spend too much resource on hiring,” he added.

GPU powers applications
To develop its applications, Xjera Labs has turned to NVIDIA GPUs instead of CPUs.

“NVIDIA GPUs compared to CPUs are much more cost effective. Using NVIDIA GPUs, our solution can support more cameras than when using CPUs. Our customers do not need to buy very expensive hardware infrastructure for their projects,” said Chu, who used CUDA during his PhD days in 2008.

CUDA is a parallel computing platform and programming model invented by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of the GPU. With CUDA, software developers can use a CUDA-enabled GPU for general purpose processing.

“We find that parallel processing has been very helpful in our research and academic pursuits. Since we were already familiar with the CUDA environment, we continued to use CUDA after we established our company,” said Chu.

Xjera Labs uses the NVIDIA Tesla P4 Inferencing Accelerator for its development work. The GPU has helped ensure that the company’s VA solutions meet customer demand for accuracy, speed, pricing, and stability.

With Xjera Labs’ solution, security is enhanced without a corresponding rise in labour cost.