Livin’ on the edge (improving video analytics efficiency)

During the last few years of my corporate career, I became involved in video analytics. While there is some overlap between video analytics and biometrics, video analytics is somewhat broader because it not only identifies individuals (via incorporation of facial recognition), but can also count people (for example, to enforce COVID capacity limits), or identify objects (for example, a particular backpack of interest that could contain an explosive device).

Because video analytics involves video rather than still images, there’s much more data that has to move from the cameras to the processing servers. For this reason, some video analytic applications take advantage of edge computing, where the analysis happens right at the edge device, removing the need to clog network bandwidth with complete video feeds.

Perhaps the edge devices only isolate the video of interest and send it off for processing. Or perhaps all of the processing takes place at the edge device.

However, as biometric and video analytics provider NEC has noted, there is a cost to edge computing.

[B]ecause cooling is difficult to manage and electricity consumption is restricted in edge devices, high-performance processors such as GPUs used in high-performance servers are not available, and processing capacity is constrained.

From https://www.nec.com/en/press/202112/global_20211201_01.html

NEC is developing a solution to address this processing capacity constraint.

Application of NEC’s newly developed gradual deep learning-based object detection technology enables efficient, high-speed, and high-precision detection of subjects from a large amount of images, even in an edge device with limited processing capacity, and enables simultaneous processing of images from multiple cameras in real time.

From https://www.nec.com/en/press/202112/global_20211201_01.html

One benefit of using software to perform the necessary calculations is that it lessens the need to upgrade hardware. As NEC and other video analytics providers well know, many organizations have already invested a lot of money in their camera systems, and would prefer software that operates with the current hardware, rather than obtaining software that requires a complete hardware replacement.

NEC’s new software isn’t available yet, but the company aims to commercialize it in 2022.

And now for the music video that is at best tangentially related to NEC’s technology advance. (And no, I don’t know if NEC’s facial recognition technology has been tested with masking of one side of the face.)

From https://www.youtube.com/watch?v=7nqcL0mjMjw