Cisco buys Modcam, NXP adds GLOW as companies smarten up edge devices with biometrics
Up and down the edge device value chain, companies are buying and building technology needed to add biometric smarts to smart cameras and other devices. Networking giant Cisco more “smarts” to its Meraki portfolio with the acquisition of Modcam, a privately held video analytics company. At the chip level, meanwhile, NXP is adding support for a popular machine learning software on a line of microcontrollers in a bid to add more biometric voice and facial recognition capabilities to cost-sensitive edge devices.
Companies and institutions trying to stitch together plans for safe return-to-work plans are looking to employ smart cameras for assistance. Facial recognition and thermal imaging are table stakes for surveillance systems now. On-device programmability and video analytics are the newest areas of focus, as recent announcements show.
Cisco announced that it acquired Modcam, a video analytics company headquartered in Malmӧ, Sweden. Modcam was founded in 2013 and had under 20 employees based on LinkedIn profiles. The company had raised $7.6 million in total funding, according to CrunchBase. Terms of the deal were not disclosed.
Cisco’s main interest in the company are the engineers with experience in machine learning and computer vision who have been developing video analytics software.
Specifically, Modcam’s technology is used for precision location and journey path analysis. Part of the unique capabilities are the software’s ability to stitch together its object detection, enabling multiple cameras to be used to piece together information on where objects are moving throughout an observed space.
One of Modcam’s original applications of its technology was foot traffic analysis for retail stores looking to understand how people make purchase decisions. Now the focus is on things like optimizing space utilization to make sure that people can safely move about an office. The other advantage to Modcam’s approach: video and analytics functions are performed on the edge device itself, offering increased data security and individual privacy.
Cisco will leverage the technology in its Meraki MV smart camera product line.
NXP adds Glow to microcontrollers
NXP cited a TIRIAS Research forecast that 98 percent of all edge devices will use some form of machine learning/artificial intelligence by 2025. NXP is busy adding those capabilities to edge devices through its i.MX RT line of microcontroller units. Microcontrollers are similar to microprocessors in that they have a central processing unit (CPU), but also have memory and communications ports integrated in a single package for cost and space advantages. These chips are typically found in embedded devices, and used to offer only a limited, dedicated functionality.
NXP’s announcement is of interest because the company is claiming to have an industry-first implementation of the Glow Neural Network machine learning software on a microcontroller. The company’s eIQ Machine Learning compiler can implement Glow models with a 2X to 3X performance gain over the standard version of Glow (which is software originally developed by Facebook). The end result: better performance of facial and voice recognition on a wider range of devices
“By using purpose-built software libraries that exploit the compute elements of their MCUs and delivering a 2-3x performance increase, NXP has demonstrated the wide-ranging benefits of using the Glow NN compiler for machine learning applications, from high-end cloud-based machines to low-cost embedded platforms,” said Dwarak Rajagopal, software engineering manager at Facebook in a prepared statement.
AI | AI chips | biometrics | biometrics at the edge | Cisco | facial recognition | machine learning | microcontroller | NXP Semiconductors | video analytics | voice recognition