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.

Download PDF version

Author profile

In case you missed it

How can security manufacturers improve after-sale service?
How can security manufacturers improve after-sale service?

What happens after the sale is complete, after the contracts are all signed and sealed? That’s when an abundance of variables can kick in – variables that can mean the difference between a successful security system or a case of buyer’s remorse. The features and value of equipment involved in a security system are well known before the sale closes, as hopefully are the integrator’s and end-user’s expectations about after-sale service. But what is the reality of after-sale service, and how can manufacturer’s make it better? We asked this week’s Expert Panel Roundtable: How can security system manufacturers improve their after-sale service for integrators and end users?

Five best practices for protecting video surveillance systems against cyber threats
Five best practices for protecting video surveillance systems against cyber threats

There’s no denying that cyber-crime is one of the biggest threats facing any organisation with the devastating results they can cause painfully explicit. Highly publicised cases stretching from the US government to digital giant Facebook has made tackling cyber security a necessity for all major organisations. The consequences of breaches have just become more severe, with new GDPR rules meaning any security breach, and resultant data loss, could cost your organisation a fine of up to four per cent of global revenue or up to 20 million euros. Cyber-crime potentially affects every connected network device. In the biggest cyber-crime to date, hackers stole $1 billion from banks around the world, by gaining access to security systems. It’s more important than ever for organisations to be vigilant when it comes to their cyber security strategy. To help avoid becoming the next victim, I’ve put together a five-point cyber plan to protect your video surveillance system. 1. Elimination of default passwords A small change to a memorable, complex password could have huge consequences for your business It is estimated that over 73,000 security cameras are available to view online right now due to default passwords. ‘Password’ and ‘123456’ are among the top five most popular passwords with a staggering 9,000,000 login details matching this description. Guessable passwords create an unsecure security system which can result in an easy way for hackers to gain access to your organisation’s data, making you vulnerable to a breach. A small change to a memorable, complex password could have huge consequences for your business. Removing default passwords from products and software forces individuals to think of their own to keep their data safe. If a password system is not provided by your organisation we recommend that your password uses two or more types of characters (letters, numbers, symbols) and it is changed periodically. 2. Encrypted firmware Encrypting firmware is an important part of any organisations overall security system. Firmware can leave an open door, allowing hackers to access your data. All firmware should be encrypted to reduce the possibilities of it being downloaded from the manufacturers website and deconstructed. If the firmware posted is not encrypted, there is a risk of it being analysed by persons with malicious intent, vulnerabilities being detected, and attacks being made. With i-PRO cameras and recorders, all firmware is securely encrypted to mitigate analysis There have been cases where a device is attacked by firmware vulnerabilities even if there are no problems with the user's settings, rendering it inoperable, and DDoS attacks being made on other servers via the device. With i-PRO cameras and recorders, all firmware is securely encrypted to mitigate analysis. There is also a possibility of being attracted to spoofing sites by targeted attack email and firmware being updated with a version that includes a virus, so firmware must always be downloaded from the vendor's page. It may also be advantageous to combine this with an imbedded Linux operating system which removes all unused features of the device, it can help to reduce the chances of malicious entities searching for backdoor entities and inserting codes. 3. Removing vulnerabilities within the operating systems Vulnerability is the name given for a functional behaviour of a product or online service that violates an implicit or explicit security policy. Vulnerabilities can occur for a number of reasons for example, due to an omission in logic, coding errors or a process failure. Network attacks exploit vulnerabilities in software coding that maybe unknown to you and the equipment provider. The vulnerability can be exploited by hackers before the vendor becomes aware. You should seek to minimise these issues by looking for a secure operating system which is regularly updated. Panasonic has developed Secure Communications, a platform and package to protect against video tampering, altering, spoofing and snooping As a provider of security solutions, Panasonic is taking a number of steps to ensure its consumers remain safe and secure. We have developed Secure Communications, a platform and package to protect against video tampering, altering, spoofing and snooping. We have combined with a leading provider of highly reliable certificates and technology for detecting and analysing cyber-attacks with its own in-house embedded cryptography technology, to provide a highly secure and robust protection layer for its embedded surveillance products. 4. Avoiding remote login using Telnet or FTP Telnet and FTP are a very outdated source of software which as a result means they lack built-in security measures Telnet and FTP are a very outdated source of software which as a result means they lack built-in security measures. File transfer protocol or transfer through cloud-based services means the files and passwords are not encrypted and can therefore be easily intercepted by hackers. An encrypted software removes the risk of files being sent to the wrong person or forwarded on without your knowledge. Telnet predates FTP and as a result is even less secure. Hyper Transfer Protocol Secure is a protocol to make secure communications by HTTP, and it makes HTTP communications on secure connections provided by SSL/TLS protocols. The major benefits of using this system is that HTTPS and VPN encrypt the communications path, so data after communications is decrypted and recorded. If recorded data is leaked, it will be in a state where it can be viewed. With data encryption, however, it remains secure and can even be recoded to storage. Thus, even if the hard drive or SD card is stolen or data on the cloud is leaked, data cannot be viewed. 5. Use of digital certificates Private and public keys are generated at manufacture in the factory and certificates installed at the factoryDigital certificates are intended to safely store the public key and the owner information of the private key it is paired with. It provides assurance that the accredited data from a third party is true and that the data is not falsified. It is beneficial for all data to be encrypted with digital certificates. Digital certificates are far safer when issued by a third party rather than creating a self-signed version unless you are 100 percent sure of the receiver identity. From April 2016, some models of Panasonic series iPro cameras come with preinstalled certificates to reduce the risk of interception and the hassle of having to create one. With i-PRO cameras with Secure function, private and public keys are generated at manufacture in the factory and certificates installed at the factory. As there is no way to obtain the private key from the camera externally, there is no risk of the private key being leaked. Also, certificates are signed by a trusted third party, and the private key used for signing is managed strictly by the authority. In addition, encryption has been cleverly implemented to reduce the usual overhead on the IP stream from 20% to 2%.

GSX to address employees' right to store firearms in vehicles
GSX to address employees' right to store firearms in vehicles

One factor aggravating concerns about workplace violence in corporate America is the easy availability of firearms. In many states, citizens, including employees, have the right to carry firearms onto a company’s property even though firearms are prohibited in the workplace. In effect, an employee prone to violence may have a firearm as near as their vehicle in the company parking lot. Currently, 23 states in the U.S. have so-called “parking lot storage” laws, which enable employees to store firearms in their vehicle’s trunk or glove box despite any corporate ban on weaponry. The laws have evolved as an expression of the Second Amendment “Right to Bear Arms” in the last decade or so. There is some variance in the laws from state to state, but they generally allow a citizen to carry a gun to and from work and keep it stored out of sight in their vehicle. Global Security Exchange (GSX) is the new branding for ASIS International’s annual conference and trade show Employee Second Amendment rights “The laws contend that employees should not have to give up their Second Amendment rights between home and the workplace and should be able to have a gun with them for protection from their front door at home to the front door of the workplace,” says Eddie Sorrells, Chief Operating Office/General Council of DSI Security Services. Sorrells will speak about the current state of the laws, how they came about, the nuances of state-to-state differences, and the possible impact on overall corporate security in a session titled “Employees Who Carry: Preventing Workplace Violence” at the upcoming GSX conference in Las Vegas, 23rd - 27th September. Global Security Exchange (GSX) is the new branding for ASIS International’s annual conference and trade show, attended by more than 22,000 security professionals from 100-plus countries. Sorrells’ session will be Sept. 24 from 10:30 to 11:45 a.m. Restrictions on gun visibility Employees may think the “right to carry” extends to the workplace, but the right only extends to the parking lot. The company still has the right to ban guns inside the premises. However, it is unlawful in some states for companies to search vehicles in parking lots, and companies who do so are violating the law, says Sorrells. Among the various state laws, some exempt public education institutions and other public venues. Depending on how the laws are written, there may be other exemptions, too. In Florida, for example, the law exempts any organisations that have explosives on site. With weapons on a company's property, a high-risk termination could potentially become violent Most laws require weapons to be stored securely out of sight. However, in Alabama, for example, it is legal to store validly permitted guns in full view during hunting season, Sorrells says. A resident of Alabama himself, Sorrells has been in the contract security business for 27 years, working mostly in multi-state operations. He has worked for 500 or so corporate security organisations throughout the country and is a practicing attorney who has studied issues of workplace violence and active shooters. “There is a political element to these laws, which were created with the goal of protecting Second Amendment rights,” says Sorrells, who says he sees arguments on both sides of the issue. However, political opinions aside, “if you’re a business owner, you have to contend with dozens of weapons on a company’s property,” he adds. “That could be an issue if a high-risk termination could potentially become violent. You have to assume there is a weapon in that person’s vehicle.” Sorrells' session will dive into the case law and illuminate some of the legal issues and how courts have addressed them GSX education program After the session, Sorrells hopes attendees will take away a good working knowledge of the state of the laws, how to comply with the laws, and issues such as posting of signs. The session will dive into the case law and illuminate some of the legal issues and how courts have addressed them. The timely session is an example of the valuable information attendees can gain by attending GSX. Sorrells has been attending the yearly ASIS International Conference and Exhibits for more than 20 years, at least since the mid-1990s. As the pre-eminent security organisation around the world, ASIS International provides unrivalled educational and networking opportunities at the yearly conference, he says. “There is a vast amount of networking and educational offerings on a wide variety of topics, including technology, legal issues, risk management, workplace violence, consulting, and anything under the sun,” says Sorrells. The newly branded GSX education program is led by subject matter experts from ASIS International, InfraGard (a public-private partnership between U.S. businesses and the Federal Bureau of Investigation), and ISSA (Information Systems Security Association). Sessions will deliver valuable, actionable takeaways to help attendees shape their security strategies. There will also be an exhibition of 550-plus suppliers and manufacturers highlighting the latest security solutions.