Marc San Pedro, Sr. Director of Software Engineering
The increasing awareness and need among both private and public sectors for security and protection on enterprise-level and individual-level has led to an explosion in the demand for video surveillance. This boost corresponds well with the technological advancements such as the availability of edge hardware, better cameras, the recent developments in signal processing, state-of-the-art big data infrastructures, new data streaming systems, AI and ML, and IoT, and has provided the thrust required for the video surveillance space to grow and flourish. However, availing intelligent video surveillance solutions is an expensive affair that requires costly hardware. Realizing this pitfall, IDEMIA National Security Solutions (IDEMIA NSS) stepped into the arena with its innovative approach and smart video products. Taking advantage of the AI algorithms and fine-tuning them for affordable devices for inference at the edge like the NVIDIA Jetson has enabled the company to deliver cutting-edge technology cost-effectively. “We bring to the table an innovative approach in how we develop our technology, we think we have good ideas, but we understand our customers know better – so we listen. We get deeply involved with our user base, our engineers will directly support missions and poc’s and tune the product in accordance to their operational needs, to address their mission and to improve outcomes. It’s the kind of feedback loop many software organizations lack.,” states Marc San Pedro, the Senior Director of Software Engineering in IDEMIA NSS.
The Identity Intelligence Video Surveillance Platform (i2VSP), the company’s flagship product, is an embedded video analytics system that leverages low-power and a high-performance computing device (GPU-accelerated parallel processing of computer vision algorithms.) The solution is capable of real-time face and pedestrian recognition, vehicle detection, color classification, and license plate recognition.
We bring to the table an innovative approach in offering our solutions by getting involved directly with the users and building the products in accordance to their needs
It can further align well with all embedded computing boards in the Jetson family, including Xavier, TX2 and Nano, providing for varied ConOps. The company serves clients on a two-pronged level—one of which is via hardware appliances that can monitor large groups of people at various choke points and the other being the nano-based solution that is aimed toward tactical covert surveillance. i2VSP also acts as an embedded video recorder (distributed VMS) that supports multiple camera streams and has a configurable storage policy. It allows a swift analysis of complex, multi-source video content, while also facilitating post-event video analysis and real-time face recognition.
What makes these unique in the core video surveillance market is its swift processing time. Its embedded, hardware-accelerated solution processes information, from the detection of a security breach to alarm generation, in well under the one-second mark. The company has also included an additional OCR capability for evaluating numbers and letters on various objects like aircraft and ships. The solution is completely autonomous, it requires no backend cloud services for any of the core capabilities: the entire pipeline executes on the edge, never leaving the device. Alongside this, the solution is also capable of generating biometric templates and matching against a central server if the customer doesn’t wish to store biometric templates on the device.
With over 40 years of experience in the surveillance industry, IDEMIA NSS has a vast outreach in law enforcement agencies, homeland security operations, and also in national defense. In light of its expertise, the company has laid great emphasis on partnering with companies manufacturing autonomous systems such as drones. It is further looking at manufacturing a payload for integration with various robotics platforms and currently working on advanced weapons detection algorithms. “We are well positioned to become the world’s leading edge analytics appliance for surveillance and security. We are incredibly passionate and proud of the overall advancement we’ve made in our edge capability over the last 3 years, our partners and customers are already planning to adopt the technology in ways that weren’t possible just a year ago.” concludes San Pedro.