Over the years, video analytics has gained an unfavourable reputation for over-promising and under-delivering in terms of performance. One of the biggest complaints regarding video analytics has been its inability to correctly identify objects in situations which appear trivial to the human observer. In many cases, this has resulted in a tendency to generate substantial numbers of false alarms, while not detecting actual events accurately. This, together with a propensity for complex set-up procedures and much need for manual fine-tuning, has prevented video analytics from becoming a mainstream application deployed on large numbers of cameras.

Machine Learning And Video Analytics

Machine Learning is a well-established field of research that has existed for decades, which is already present in many products and applications. Machine Learning is based on collecting large amounts of data specific to a particular problem, training a model using this data and then employing this model to process new data. With regard to video analytics, one of the most critical problems impacting accuracy is object classification. Fundamental to improving performance is the capability to teach the algorithm to distinguish between people, animals, different types of vehicles and sources of noise at an extremely high level of accuracy.

Deep Learning Algorithms

The recent increased interest in Deep Learning is largely due to the availability of graphical processing units

Until recently there have been minimal applications of Machine Learning used in video analytics products, largely due to high complexity and high resource usage, which made such products too costly for mainstream deployment. However, the last couple of years have seen a tremendous surge in research and advances surrounding a branch of Machine Learning called Deep Learning. Deep Learning is a name used to describe a family of algorithms based on the concept of neural networks. Very loosely speaking, these algorithms try to emulate the functionality of the brain’s neurons, enabling them to learn efficiently from example, and subsequently apply this learning to new data.

The recent increased interest in Deep Learning is largely due to the availability of graphical processing units (GPUs). GPUs can efficiently train and run Deep Learning algorithms, and have allowed the scientific community to accelerate their research and application, bringing them to the point where they exceed the performance of most traditional Machine Learning algorithms across several categories.

Solving Object Classification With Deep Learning

This means that Deep Learning can now be used to solve the most crucial problem facing video analytics – object classification – by collecting many thousands of images from hundreds of surveillance cameras, which must first be manually labelled and classified by a human, into a range of categories that include: person, car, bus, truck, bird, vegetation, dog and many more. To achieve the required accuracy rates, such a vast database must be collected and identified from actual surveillance footage.

A crucial component in achieving and maintaining the high performance of Deep Learning-based applications is the ability to continuously update the models as more data is collected
Basic classification and false alarm reduction are the first applications of Deep Learning for video analytics

A Deep Learning algorithm trained on images collected from YouTube, Google Search and elsewhere on the Internet will completely fail in analysing images from surveillance cameras, due to the difference in viewing angles, resolution and image quality. Once enough images are collected, a Deep Learning classifier algorithm can be trained and deployed as part of a video analytics solution, enabling it to practically eliminate most of the existing causes for false alarms.

Due to GPU requirements in order for the algorithms to run efficiently, video analytics solutions using Deep Learning will initially need to run on a server. A few solutions of this nature are already available and are showing a dramatic leap in performance in comparison to traditional video analytics, with a drastic reduction in false alarm rates and a significant increase in detection accuracy. Concurrently, these new solutions do not require manual tweaking by the user and are essentially plug-and-play, making mass deployment a realistic premise.

Surveillance Applications Of Deep Learning

Basic classification and false alarm reduction are the first applications of Deep Learning for video analytics, but they are by no means the only ones. In the not-too-distant future, we will see Deep Learning enabling as yet not possible video analytics applications, such as identifying objects carried by people, such as a gun, handbag, or a knife, or being able to quickly find people and vehicles with similar appearances across multiple cameras and more.

Over the next few years, we will see a transition of video analytics using Deep Learning running on servers to running inside cameras

Over the next few years, we will see a transition of video analytics using Deep Learning running on servers, to running inside cameras, as powerful, low-cost hardware capable of running Deep Learning becomes more available and a basic function of newer surveillance camera models. This will push the acceptance of video analytics even further, eventually making it a fundamental element of every surveillance camera deployed.

Increasing Accuracy With Updates

A crucial component in achieving and maintaining the high performance of Deep Learning-based applications, is the ability to continuously update the models as more data is collected so that the models increase in accuracy. This will give an advantage to cloud-based video analytics services, since they can collect vast amounts of data from cameras connected to the service, train new models in the cloud based on this data and then push these new models to cameras at the edge. This continuous improvement cycle will be instrumental in helping video analytics fulfil the promise of improving peoples’ safety and security, by giving surveillance cameras human-level accuracy and a comprehensive understanding of the environment.

Share with LinkedIn Share with Twitter Share with Facebook Share with Facebook
Download PDF version

Author profile

In case you missed it

Motorola Solutions Acquires Pelco For $110 Million
Motorola Solutions Acquires Pelco For $110 Million

Motorola Solutions, Inc. today announced it has completed the acquisition of Pelco, Inc., a global provider of video security solutions based in Fresno, California. Pelco designs, develops and distributes end-to-end video technology, including video security cameras and video management system software. The company’s scalable solutions and commitment to service delivery enables customers of all sizes to mitigate risk, increase operational efficiencies and enhance safety.  “Video continues to play a more powerful role in enabling safer cities and securing businesses around the world,” said Greg Brown, chairman and CEO, Motorola Solutions. “Pelco’s track record of innovation, internationally recognized brand, global channel and customer installed base enable us to further expand our global footprint with enterprise and public safety customers.”

What are the Security Challenges of Protecting the Cannabis Industry?
What are the Security Challenges of Protecting the Cannabis Industry?

The advent of a truly new market for the physical security industry is a rare occurrence. Particularly rare is a new market that is both fast-growing and provides an environment that is not just conducive to application of physical security technologies but that actually demands it. Such is the case with the market for legalized marijuana. We asked this week’s Expert Panel Roundtable: What are the security challenges of protecting the cannabis industry?

Wireless Technology Is Transforming Motion Detection
Wireless Technology Is Transforming Motion Detection

Motion detection is a key feature of security systems in residential and commercial environments. Until recently, systems have relied heavily on closed circuit television (CCTV) and passive infrared (PIR) sensors, which both require significant investment and infrastructure to install and monitor. Developments in wireless technology are increasing home security possibilities. Few years ago, these developments led Cognitive Systems to discover that the wireless signals surrounding oneself can be used to detect motion. Known in the wireless industry as WiFi sensing, this technology brings many benefits that other motion detection solutions have not been able to provide. The working of WiFi sensing At Cognitive Systems, the company has used WiFi sensing technology to develop a motion detection solution called WiFi Motion™, which measures and interprets disruptions in RF signals transmitted between WiFi devices. When movement occurs in a space, ripples in the wireless signals are created. WiFi Motion interprets these ripples and determines if an action, such as sending a notification, is needed. Enabling this functionality in a space is incredibly simple. With a software upgrade to only one’s WiFi access point (or mesh router), motion sensing capabilities are layered into one’s WiFi network. Existing connected WiFi devices then become motion detectors without detracting from their original functions or slowing down the network. Using artificial intelligence (AI), WiFi Motion establishes a benchmark of the motionless environment and learns movement patterns over time, which could be used to predict trends. This allows unusual movement patterns to be detected with greater accuracy while decreasing the potential for costly false alerts. WiFi Motion requires no line-of-sight or installation WiFi sensing and other home monitoring solutions All of these capabilities are made possible by WiFi sensing and together create a motion detection system that provides unparalleled accuracy, coverage, privacy and affordability compared to other solutions on the market. PIR integration is far more complex and imposes electronic and physical design restrictions compared to WiFi sensing. In terms of placement, PIR systems are difficult to install, requiring line-of-sight and a device in every room for localization. WiFi Motion requires no line-of-sight or installation and is also a scalable solution compared to PIR. Much like cameras, PIRs can only cover so much space, but WiFi Motion can cover the entire home and even detect motion in the dark and through walls, without adding additional devices to the home. WiFi Motion detects less distinguishing context than cameras and microphones, but more context than regular PIR sensors for the perfect balance of privacy and highly accurate motion detection. Privacy solution While cameras have been the security solution for years, WiFi Motion offers a more affordable solution that can rival the privacy and coverage capabilities of even the most high-end cameras. With such a wide coverage area, one might think that WiFi sensing infringes on privacy, but actually, the opposite is true. With WiFi Motion, the contextual information collected cannot be used to identify a specific individual, unlike cameras which can clearly identify a person’s face or microphones, which can identify a person’s voice. It is different from other smart home security options that use cameras and microphones because it only senses motion using WiFi signals - it doesn’t “see” or “listen” like a camera or microphone would. This provides opportunities for added security in spaces where privacy might be a concern and installing a camera may not be a comfortable solution, such as bathrooms and bedrooms. The data collected is also anonymized and highly encrypted according to stringent industry privacy standards. Existing connected WiFi devices then become motion detectors Additional WiFi sensing applications Since WiFi sensing technology requires no additional hardware or subscription fees, it is much more affordable than other motion detection solutions. It can be used as a standalone solution, or it can be easily layered into more complex systems. This ease of integration, scalability and relatively low cost brings a lot of potential for various applications. Motion detection can trigger other smart devices in the network to turn lights on or off In eldercare, for example, WiFi sensing can be used to help seniors live comfortably in their homes for as long as possible. With the increasing aging population and high costs associated with care homes, the market for this application is considerable. Caregivers can use an app to monitor movement in their loved one’s home and be alerted about unusual movement patterns that could indicate a concern. For smart homes and other environments that have a network of smart devices, the artificial intelligence (AI) component of the technology allows for improvements to automated features. Motion detection can trigger other smart devices in the network to turn lights on or off or make adjustments to the temperature in a room. Security for the commercial sector For office buildings and other commercial properties, it is easy to see how all of these features could be scaled up to offer a highly accurate and cost-effective motion sensing and smart device automation solution. Cognitive Systems is closely involved with the development of WiFi sensing technology, working with various industry groups to establish standards and help it reach its full potential. WiFi Motion is merely the tip of the iceberg in terms of motion sensing possibilities, but its applications in the world of security are undeniably compelling. It is an exciting time for the wireless industry, as one works with stakeholders in the security space to explore everything this technology can do.