Deep Learning in Surveillance
At the Edge

Deep Learning in Surveillance

As the surveillance industry gathers at IFSEC London next week, some of the trends we expect to see are around deep learning and the actionable insights surveillance data can provide- at the edge, on NVRs, or in the cloud.

The concept of analyzing video to extract valuable insights is at the heart of deep learning. Effective deep learning in surveillance requires both computing power and many hours of “training” video, which is actual surveillance video footage from which objects and behavioral patterns can be analyzed. Deep learning ultimately leads to very competent artificial intelligence (AI), but can require thousands of hours of video footage to discern one human behavioral pattern. The opportunity to extract actionable insights from surveillance video is driving users to hold onto more footage, and hold onto it longer, stoking data storage growth even further. Deep learning continues to promise new ways to use surveillance video, to go beyond the traditional surveillance use-case and drive the industry to capture more, store more, and analyze more surveillance video.

Deep Learning at the Surveillance Edge

Deep learning is finding its way closer to the surveillance edge. Cameras are being outfitted with more capable edge computing, allowing cameras to autonomously perform analysis without sending video to an upstream server or AI-enabled NVR. New technology allows cameras to scan an on-camera database of thousands of face images to automatically detect persons of interest in public gathering places and alert authorities in real time. Cameras will be able to analyze flows of people to alert staff of growing queues, or detect if someone has fallen or collapsed and alert medical staff. All of these will be done on the camera itself in real time, without relying on a back-end server or the cloud which consumes both bandwidth and time.

Deep Learning Adding Intelligence to NVRs

A second trend is the expanding intelligence of NVRs and VMS servers with deep learning capabilities. Where cameras are not yet ready for a broader role as edge computing devices, near-edge deep learning is taking place on surveillance video edge servers, whether on server-based VMS or AI-enabled NVRs. VMS solutions are being architected or optimized to best use new GPUs and CPUs to improve overall deep learning capability, and speed algorithms related to object recognition and facial recognition. NVRs with deep learning take advantage of the greater storage capacity and potentially more sophisticated processing versus individual cameras to perform more advanced analytics, such as locating an individual face image from weeks or months of stored video, or creating traffic heat maps from hours of retail surveillance video.

VSaaS Cloud Storage Enables Data Retention for Deep Learning

A third trend is the establishment of more dedicated cloud storage for surveillance video.  VSaaS has been recognized as a fast growing offering, and surveillance video cloud storage offers two advantages – one is to relieve on-site storage costs and management while still retaining the large amount of video data needed to extract valuable insights; the other is to pool mass amounts of surveillance video close to the strong analytics server or cloud infrastructure. With the recognition that deep learning amplifies the potential value extracted from surveillance video, it’s no wonder that more surveillance video is being retained, and the cloud represents a real option due to improved scalability and cost.

A Portfolio for Storage at the Edge, NVR or Cloud

For effective deep learning to happen at the camera, at the NVR, or in the cloud, more and more surveillance video must be stored, and must be stored close to where deep learning happens. That’s where Western Digital comes in. We have an unmatched portfolio of storage devices for surveillance cameras; for DVRs, NVRs and server-based VMSs; and for the analytics back-end server/cloud.  We’re excited by what deep learning brings to the industry, and are excited that our technology can help improve the effectiveness of deep learning, wherever it’s taking place.

IFSEC Preview: Partners in Edge, NVR and Cloud Surveillance

Western Digital will be at IFSEC International in London, June 19-21, to showcase our edge-to-core storage technologies for surveillance. We’ve also partnered with pioneers who continue to advance innovation, in ways that we capture, store, and analyze surveillance video. Visit us in stand #E200 to see how we’re working with these leaders to continue to drive beyond surveillance to valuable new ways to leverage surveillance. Hear about how Axis Communications uses edge storage to provide decentralized video recording and maintain high resolution recording in deployments with bandwidth limitations. See our surveillance storage optimized for Hikvision’s DeepInMind NVRs, able to handle a high number of simultaneous streams while reducing frame loss and image pixelation. And see how innovative analytics solutions can be implemented via Milestone’s XProtect solutions using our enterprise-class storage utilized in cloud-based storage infrastructures.

A Look Ahead: Deep Learning in Surveillance

Surveillance has come a long way. No longer are we using surveillance video to see what happened. We’re seeing use of deep learning on surveillance video to predict what will happen.  The future potential of surveillance video comes from extracting actionable insights, and we’re excited to provide storage solutions to go beyond surveillance.

Forward-Looking Statements

Certain blog and other posts on this website may contain forward-looking statements, including statements relating to expectations for our product portfolio, the market for our products, product development efforts, and the capacities, capabilities and applications of our products. These forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially from those expressed in the forward-looking statements, including development challenges or delays, supply chain and logistics issues, changes in markets, demand, global economic conditions and other risks and uncertainties listed in Western Digital Corporation’s most recent quarterly and annual reports filed with the Securities and Exchange Commission, to which your attention is directed. Readers are cautioned not to place undue reliance on these forward-looking statements and we undertake no obligation to update these forward-looking statements to reflect subsequent events or circumstances.

 

Leave a Response

Brian Mallari
Brian Mallari is a Director of Product Marketing for surveillance and enterprise products at Western Digital Corporation. In this position he is responsible for product strategy and execution of the company's surveillance and enterprise HDD business objectives.