The field of artificial intelligence known as machine learning or cognitive computing has in recent years become highly popular. The meteoric rise of “deep learning” technology over the past several years has been truly dramatic in many industries. Industry giants from Google, Microsoft, Facebook, IBM and many others have been pouring massive amounts of investments in this field of artificial intelligence. The machine learning field has exploded on the scene with the breakthrough in the new “deep learning” technology.

Developments in deep learning have ramifications for the physical security industry, too. In video analytics, for example, deep learning has shown promise to improve some difficult problems, although more work is needed. This article will cover the evolving field of deep learning and its potential impact on the security and video surveillance markets.

Evolution Of Deep Learning

The field of deep learning evolved from “artificial neural nets” from the 1980s. In the early years of this branch of artificial intelligence, the neural nets are modeled after a human’s brain, which consists of over 100 billion neurons. The field of neural networks never really took off in the ‘80s and ‘90s due to many reasons. The key limitations of the earlier systems are the difficultly to train the network; and the hardware CPU technologies were too slow to properly train a neural net that can solve meaningful real-world applications.

Over the past several years, real
world applications of deep learning
now encompass many industries
including handwriting recognition
and language translation

The 1980s and 1990s were the dark days of neural network research. Since 2000, the research community of neural nets has really started to garner industry labs’ attention from the breakthroughs in deep learning work in academia at the University of Toronto, NYU, Stanford and others. Over the past several years, real world applications of deep learning now encompass many industries including handwriting recognition, language translation, automatic game (chess/Go) playing, object classification, face recognition, medical image analysis, autonomous driving cars and many other fields.

One example of the excitement with deep learning technology is the recent breakthrough from Google’s AlphaGo, a computer program that for the first time beat a professional human Go player in October, 2015. The sophistication of the Deep learning based program has astonished many in the field of artificial intelligence due the complexity of the ancient Asia GO game, which is considerably more complicated than chess.

 Video Surveillance Applications For Deep Learning

Although deep learning has been applied to many industries with breakthrough results compared to legacy systems, not all applications are suitable for deep learning. In the field of video surveillance, several applications stand out that can benefit from deep learning.

Face recognition. Deep learning technology has significantly improved the accuracy rate of face recognition. The National Institute of Standards and Technology (NIST) has conducted Face Recognition Vendor Test (FRVT) test over the past decade. The improvements over the last 20 years of face recognition error rates have decreased by three orders of magnitude, according to an NIST Interagency Report. Most of today’s top-performing commercial face recognition products are based on deep learning. The accuracy has reached 99.9% for controlled environments like airport immigration face recognition applications, according to research by Facebook and Tel Aviv University.

Person and object detection. Person detection and object detection is another area where deep learning has shown tremendous progress. For example, over the past five years, the IMAGENET database has organized the “large scale visual recognition challenge,” in which image software algorithms are challenged to detect, classify and localize a database of over 150,000 photographs collected from Flickr and other search engines. The dataset is labeled into 1,000 object categories. Many deep learning systems are trained with over 1.2 million images from the IMAGENET dataset running on GPU based hardware accelerators. The improvements in accuracy range from 72% to over 90% from 2010 thru 2014. In 2015, all IMAGENET contestants used deep learning techniques.

Deep learning in its infancy has shown a lot of promise in improving some difficult video analytics problems
In the field of video surveillance, several applications stand out that can benefit from deep learning

Deep Learning-Based Video Surveillance Solutions

A key advantage of deep learning-based algorithms over legacy computer vision algorithms is that deep learning system can be continuously trained and improved with better and more datasets. Many applications have shown that deep learning systems can “learn” to achieve 99.9% accuracy for certain tasks, in contrast to rigid computer algorithms where it is very difficult to improve a system past 95% accuracy.

Deep learning has the true potential
of significantly reducing false-positive
detection events that plague many
security video analytics systems

The second advantage with deep learning system is the “abnormal” event detection. Deep learning systems have shown remarkable ability to detect undefined or unexpected events. This feature has the true potential of significantly reducing false-positive detection events that plague many security video analytics systems. In fact, the inability to reduce false-positive detection rate is the key problem in video surveillance industry; and has to-date prevented the wide scale acceptance of many vendor’s intelligent video analytics solutions.

Open Issues With Deep Learning Technology For Video

Deep learning in its infancy has shown a lot of promise in improving some hard, and difficult video analytics problems. Much more work needs to be done to fine-tune the generic deep learning system to learn and detect domain specific events that are unique to security-oriented environments.

The second challenge is that the engineering talent for deep learning is in extreme short supply. Most graduates today come from the top universities and upon graduation are immediately snapped up by Internet giants like Google, Facebook, Amazon, Microsoft, etc. The competition for trained machine learning engineers is intense.

The third challenge is that not all video analytics algorithms are best applied with deep learning. There are many legacy computer vision algorithms that have been developed over decades that are very well suited and deployed in commercially successful products. For example, license plate recognition performs very well with computer-vision based algorithms. Industry needs to do more research in hybrid systems the combine the best of computer vision algorithms and deep learning.

What’s Next?

Similar to cloud computing and big data technologies, deep learning technology is now emerging as the third wave of rapid advances that have taken over the information industry by storm.

Over the next decade, very few areas in the technology sector will not be touched by the advances of cloud, big data and machine learning. For the video surveillance industry, this is welcome news. The industry has been lacking in innovations that can significantly advance the state-of-the art.

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

Author profile

In case you missed it

Water Plant Attack Emphasizes Cyber’s Impact On Physical Security
Water Plant Attack Emphasizes Cyber’s Impact On Physical Security

At an Oldsmar, Fla., water treatment facility on Feb. 5, an operator watched a computer screen as someone remotely accessed the system monitoring the water supply and increased the amount of sodium hydroxide from 100 parts per million to 11,100 parts per million. The chemical, also known as lye, is used in small concentrations to control acidity in the water. In larger concentrations, the compound is poisonous – the same corrosive chemical used to eat away at clogged drains. The impact of cybersecurity attacks The incident is the latest example of how cybersecurity attacks can translate into real-world, physical security consequences – even deadly ones.Cybersecurity attacks on small municipal water systems have been a concern among security professionals for years. The computer system was set up to allow remote access only to authorized users. The source of the unauthorized access is unknown. However, the attacker was only in the system for 3 to 5 minutes, and an operator corrected the concentration back to 100 parts per million soon after. It would have taken a day or more for contaminated water to enter the system. In the end, the city’s water supply was not affected. There were other safeguards in place that would have prevented contaminated water from entering the city’s water supply, which serves around 15,000 residents. The remote access used for the attack was disabled pending an investigation by the FBI, Secret Service and Pinellas County Sheriff’s Office. On Feb. 2, a compilation of breached usernames and passwords, known as COMB for “Compilation of Many Breaches,” was leaked online. COMB contains 3.2 billion unique email/password pairs. It was later discovered that the breach included the credentials for the Oldsmar water plant. Water plant attacks feared for years Cybersecurity attacks on small municipal water systems have been a concern among security professionals for years. Florida’s Sen. Marco Rubio tweeted that the attempt to poison the water supply should be treated as a “matter of national security.” “The incident at the Oldsmar water treatment plant is a reminder that our nation’s critical infrastructure is continually at risk; not only from nation-state attackers, but also from malicious actors with unknown motives and goals,” comments Mieng Lim, VP of Product Management at Digital Defense Inc., a provider of vulnerability management and threat assessment solutions.The attack on Oldsmar’s water treatment system shows how critical national infrastructure is increasingly becoming a target for hackers as organizations bring systems online “Our dependency on critical infrastructure – power grids, utilities, water supplies, communications, financial services, emergency services, etc. – on a daily basis emphasizes the need to ensure the systems are defended against any adversary,” Mieng Lim adds. “Proactive security measures are crucial to safeguard critical infrastructure systems when perimeter defenses have been compromised or circumvented. We have to get back to the basics – re-evaluate and rebuild security protections from the ground up.” "This event reinforces the increasing need to authenticate not only users, but the devices and machine identities that are authorized to connect to an organization's network,” adds Chris Hickman, Chief Security Officer at digital identity security vendor Keyfactor. “If your only line of protection is user authentication, it will be compromised. It's not necessarily about who connects to the system, but what that user can access once they're inside. "If the network could have authenticated the validity of the device connecting to the network, the connection would have failed because hackers rarely have possession of authorized devices. This and other cases of hijacked user credentials can be limited or mitigated if devices are issued strong, crypto-derived, unique credentials like a digital certificate. In this case, it looks like the network had trust in the user credential but not in the validity of the device itself. Unfortunately, this kind of scenario is what can happen when zero trust is your end state, not your beginning point." “The attack on Oldsmar’s water treatment system shows how critical national infrastructure is increasingly becoming a target for hackers as organizations bring systems online for the first time as part of digital transformation projects,” says Gareth Williams, Vice President - Secure Communications & Information Systems, Thales UK. “While the move towards greater automation and connected switches and control systems brings unprecedented opportunities, it is not without risk, as anything that is brought online immediately becomes a target to be hacked.” Operational technology to mitigate attacks Williams advises organizations to approach Operational Technology as its own entity and put in place procedures that mitigate against the impact of an attack that could ultimately cost lives. This means understanding what is connected, who has access to it and what else might be at risk should that system be compromised, he says. “Once that is established, they can secure access through protocols like access management and fail-safe systems.”  “The cyberattack against the water supply in Oldsmar should come as a wakeup call,” says Saryu Nayyar, CEO, Gurucul.  “Cybersecurity professionals have been talking about infrastructure vulnerabilities for years, detailing the potential for attacks like this, and this is a near perfect example of what we have been warning about,” she says.  Although this attack was not successful, there is little doubt a skilled attacker could execute a similar infrastructure attack with more destructive results, says Nayyar. Organizations tasked with operating and protecting critical public infrastructure must assume the worst and take more serious measures to protect their environments, she advises. Fortunately, there were backup systems in place in Oldsmar. What could have been a tragedy instead became a cautionary tale. Both physical security and cybersecurity professionals should pay attention.

Expert Roundup: Healthy Buildings, Blockchain, AI, Skilled Workers, And More
Expert Roundup: Healthy Buildings, Blockchain, AI, Skilled Workers, And More

Our Expert Panel Roundtable is an opinionated group. However, for a variety of reasons, we are sometimes guilty of not publishing their musings in a timely manner. At the end of 2020, we came across several interesting comments among those that were previously unpublished. Following is a catch-all collection of those responses, addressing some of the most current and important issues in the security marketplace in 2021.

Smart Offices: How Is Mobile ID Changing The Way We Access The Office?
Smart Offices: How Is Mobile ID Changing The Way We Access The Office?

If you’re a security or facilities manager, you may already be aware of the quiet revolution that’s taking place across businesses and organizations up and down the country. By the end of 2020, 20% of all ID and access control systems featured mobile capability, and this is set to increase by a further 34% over the next three years. There’s no doubt that using a smartphone or mobile device in place of traditional credential and access control is a growing trend that’s only been sped up by the pandemic. It’s true that many businesses are still very much focused on remote working, although many are now starting to implement new-and-improved strategies that are better suited to protect the workforce moving forward. Mobile ID systems As the next normal becomes clearer, businesses will be reviewing procedures such as access control, occupancy monitoring, reducing touch points, and tracking visitors. Mobile ID systems are ideally suited to this task. But what are the key reasons for considering such a setup in 2021? But why is this new technology so well-suited to future-proof your physical access system, and why is it becoming so popular? Eradicating outdated legacy credentials Have you seen just how vulnerable outdated Proximity card technology can be? Low-frequency 125kHz cards can be cloned in a matter of seconds with the use of cheap, readily available tools. Despite their weaknesses, they are still used by a huge majority of businesses – big and small. All smartphones include two industry-standard features that make them perfect for operating a secure, contactless credential Replacing such a system with a mobile-enabled system is one of the best ways to increase security ten-fold. Thanks to a cloud-based infrastructure, mobile ID offers best-in-class security and cryptography. All smartphones include two industry-standard features that make them perfect for operating a secure, contactless credential. Bluetooth Smart and NFC (Near Field Communication) make them the best product to operate such a credential via a secure app. If you’re looking for best-in-class security in 2021, mobile access is most definitely the way forward. Removing touch points across the business Reducing touch points and the adoption of touchless facilities has become a key priority for businesses in the wake of COVID-19. Even as businesses start to return to the office and operate a home/office split, it will be imperative that unnecessary contact is kept to an absolute minimum between staff. The traditional issuance of identification and access control credentials can pose problems in this regard. Facility and security managers who are responsible for onboarding and processing ID have done the process face to face. Mobile access makes it possible to carry this process out without people coming into direct content. First, the security manager has access to a secure portal, allowing them to create, manage and edit credentials anywhere. They can upload and remotely transfer mobile ID and access control credentials directly to users’ smartphones over the air. Via the secure app, users can view and see their credentials and immediately begin using it for ID and access control by simply placing their smartphone over card readers. Enabling a more flexible way of working The way in which we work has changed for good. Even as people more people return to the office in 2021, a majority of businesses will be operating a home/office split indefinitely. This once again reinforces the need for a smarter, more adaptable onboarding system. Implementing mobile ID is the perfect way of doing this: over-the-air delivery of credentials and security data is now a given, helping businesses create the perfect balance between the home and the office. No longer do people have to come into the office for the onboarding process. Increasing convenience and user experience More often businesses are realising the value mobile ID can have for enhancing the work experience as well as security Ok, so mobile ID is the perfect way of increasing security and adapting workplaces to a post-COVID way of working. And we’ve not even touched on the most obvious advantage yet: Convenience. How many times have you forgotten your ID card? We’re sure it’s more times than you forget your smartphone. These powerful processors have become intertwined with the way we carry out tasks on a daily basis. They’re so vital that people will soon notice if they’ve forgotten it. From an employee’s perspective, mobile ID and access control is simple, convenient, and extremely user-friendly. More and more businesses are realizing the value mobile ID can have for enhancing the work experience as well as security. From the employer’s perspective, mobile ID means it’s easier for administrators to manage access and credentials. Future-proofing access control now will ensure that in the longer term, mobile ID is well worth the investment. The annual expenditure of printing ID cards and purchasing credentials can be vast, while reissuance costs can also quickly add up for larger organizations. These issues are a thing of the past for businesses using mobile ID. Mobile ID perfect tool for 2021 and beyond Until mobile ID, new and improved credentials’ main focus was on increasing security. Mobile ID not only delivers that, but it also provides a more convenient way of accessing the office in a way that’s perfectly suited to returning to the office in 2021. If there was ever a time to upgrade, now is the time. Summing up, mobile access is changing the way we access the office by: Eliminating weak links in security systems such as outdated legacy card technologies Eradicating the need for touch points across multiple areas of the workplace Enabling a smarter, more flexible approach to onboarding Increasing convenience – for both employers and employees.