Data storage devices across the security industry are routinely required to handle an enormous amount and many layers of raw data. As Safe City projects in varying sizes become more prevalent, the number of surveillance nodes has reached the hundreds of thousands. And due to the widespread use of high-definition monitoring, the amount of data involved in security surveillance has increased dramatically in a short time. Efficient collection, analysis, and application of data and the intelligent use of it are becoming ever more critical in this industry. Thus, improving video intelligence appears to be an inevitable, industry-wide goal.

Security users hope that their investment in new products will bring even more benefits beyond simply tracing and tracking persons of interest and evidence collection after a security event. Some examples of added benefits include using the latest technologies to replace the large amount of man-power previously required for searching surveillance footage, detecting anomalous data, and finding ever more efficient ways to allow surveillance to shift from post-incident tracing to alerts during incidents—or even pre-incident alerts. In order to satisfy these demands, new technologies are required. Intelligent video surveillance has been available for many years. However, the outcomes of its application have not been ideal. The emergence of deep learning has enabled these demands to become reality.

The insufficiency of traditional intelligent algorithms

Traditional intelligent video surveillance has especially strict requirements for a scene’s background. The accuracy of intelligent recognition and analysis in comparable scenarios remains inconsistent. This is primarily due to the fact that traditional intelligent video analysis algorithms still have many flaws. In an intelligent recognition and analysis process, such as human facial recognition, two key steps are required: First, features are extracted, and second, “classification learning” is performed.

The features in traditional intelligent algorithms are designed by humans and have always been heavily subjective

The degree of accuracy in this first step directly determines the accuracy of the algorithm. In fact, most of the system’s calculation and testing workload is consumed in this part. The features in traditional intelligent algorithms are designed by humans and have always been heavily subjective. More abstract features—those that humans have difficulty comprehending or describing—are inevitably missed. With shifting angles and lighting, and especially when the sample size is enormous, many features can be too difficult to detect. Therefore, while traditional intelligent algorithms perform well in very specific environments, subtle changes (image quality, environment, etc.) yield significant challenges to accuracy.

Target detection and attribute recognition

The second step—classification learning—mainly involves target detection and attribute recognition. As the number of available categories for classification rises, so does the difficulty level. Hence, traditional intelligent analysis technologies are highly accurate in vehicle analysis but not in human and object analysis. For example, in vehicle detection, a distinction is made between a vehicle and a non-vehicle, so the classification is simple and the level of difficulty is low. To recognise vehicle attributes requires recognition of different vehicle designs, logos, and so on. However, there are relatively few of these, making the classification results generally accurate. On the other hand, if recognition is to be performed on human faces, each person is a classification of its own, and the corresponding categories will be extremely numerous—naturally leading to a very high level of difficulty.

The accuracy of intelligent recognition and analysis in comparable scenarios remains inconsistent
Enhanced accuracy is the result of multi-layer learning and extensive data collection

Traditional intelligent algorithms generally use shallow learning models to handle situations with large amounts of data in complex classifications. The analysis results are far from ideal. Furthermore, these results directly restrict the breadth and depth of intelligent applications and further development. Hence the need for increasing the “depth” of intelligence in big data for the security industry is arising.

The advantages of Deep Learning and its algorithms

Traditional intelligent algorithms are designed by humans. Whether or not they are designed well depends greatly on experience and even luck, and this process requires a lot of time. So, is it even possible to get machines to automatically learn some of the features? Yes! This is actually the objective of Artificial Intelligence (AI).

The inspiration for deep learning comes from a human brain’s neural networks. Our brains can be seen as a very complex deep learning model. Brain neural networks are comprised of billions of interconnected neurons; deep learning simulates this structure. These multi-layer networks can collect information and perform corresponding actions. They also possess the ability for object abstraction and recreation. Deep learning is intrinsically different from other algorithms. The way it solves the insufficiencies of traditional algorithms is encompassed in the following aspects.

Algorithmic model for deep learning

The algorithmic model for deep learning has a much deeper structure than the two 3-layered structures of traditional algorithms. Sometimes, the number of layers can reach over a hundred, enabling it to process large amounts of data in complex classifications. Deep learning is very similar to the human learning process, and has a layer-by-layer feature-abstraction process. Each layer will have different “weighting,” and this weighting reflects on what was learned about the images’ “components.” The higher the layer level, the more specific the components. Simulating the human brain, an original signal in deep learning passes through layers of processing; next, it takes a partial understanding (shallow) to an overall abstraction (deep) where we can perceive the object.

Deep learning does not require manual intervention but relies on a computer to extract features by itself. This way it is able to extract as many features from the target as possible, including abstract features that are difficult or impossible to describe. The more features there are, the more accurate the recognition and classification will be. Some of the most direct benefits that deep learning algorithms can bring include achieving comparable or even better-than-human pattern recognition accuracy, strong anti-interference capabilities, and the ability to classify and recognise thousands of features.

Hikvision has operated in the security industry for many years with its own research and development capabilities

Key factors of Deep Learning

In total, there are three main reasons why deep learning only became popular in recent years and not earlier: the scale of data involved, computing power, and network architecture.

Improvements in data-driven algorithm performance have accelerated deep learning in various intelligent applications in a short amount of time. Specifically, with the increase in data scale, algorithmic performance improved as well. Accordingly, user experience has improved and more users are involved, further facilitating a larger scale of data.

Video surveillance data makes up 60% of big data, and the amount is rising at 20% annually. The speed and scale of this achievement is due to the popularisation of high definition video surveillance—HD 1080p is becoming more common, and 4K and higher resolutions are gradually being applied in many important applications.

Hikvision has operated in the security industry for many years with its own research and development capabilities, employing large amounts of real video and image data as training samples. With a large amount of good quality data, and over a hundred team members to label the video images, sample data with millions of categories have been accumulated. With this large amount of quality training data, human, vehicle, and object pattern recognition models will become more and more accurate for video surveillance use.

The deep learning model requires a large amount of samples, making a large amount of calculations inevitable
Enhanced accuracy is the result of multi-layer learning and extensive data collection

Higher computational power

Furthermore, high performance hardware platforms enable higher computational power. The deep learning model requires a large amount of samples, making a large amount of calculations inevitable. In the past, hardware devices were incapable of processing complex deep learning models with over a hundred layers. In 2011, Google’s DeepMind used 1,000 devices with 16,000 CPUs to simulate a neural network with approximately 1 billion neurons. Today, only a few GPUs are required to achieve the same sort of computational power with even faster iteration. The rapid development of GPUs, supercomputers, cloud computing, and other high performance hardware platforms has allowed deep learning to become possible.

Finally, the network architecture plays its own role in advancing deep learning. Through constant optimisation of deep learning algorithms, better target-object recognition can be achieved. For more complex applications such as facial recognition or in scenarios with different lighting, angles, postures, expressions, accessories, resolutions, etc., network architecture will impact the accuracy of recognition, i.e., the more layers in deep learning algorithms, the better the performance.

In 2016, Hikvision achieved the number one position in the Scene Classification category at the ImageNet Large Scale Visual Recognition Challenge 2016. The team from Hikvision Research Institute used inception-style networks and not-so-deep residual networks that perform better in considerably less training time, according to Hikvision’s experiments for training and testing.

Furthermore, Hikvision’s Optical Character Recognition (OCR) Technology, based on Deep Learning and led by the company’s Research Institute, also won the first price in the ICDAR 2016 Robust Reading Competition. The Hikvision team substantially surpassed both strong domestic and foreign competitors in three word-recognition challenges, including born-digital images, focused scene text, and incidental scene text, demonstrating that the word recognition technology by Hikvision reached the world’s top level.

In the past two years, deep learning technology has excelled in speech recognition, computer
vision, voice translation,
and much more

Application of Deep Learning products

In the past two years, deep learning technology has excelled in speech recognition, computer vision, voice translation, and much more. It has even surpassed human capabilities in the areas of facial verification and image classification; hence, it has been highly regarded in the field of video surveillance for the security industry.

In the application of intelligent video in target detection, tracking, and recognition, the rise of deep learning has had a profound influence. When applying those three functions, deep learning potentially touches upon every aspect of the security video surveillance industry: facial detection, vehicle detection, non-motor vehicle detection, facial recognition, vehicle brand recognition, pedestrian detection, human body feature detection, abnormal facial detection, crowd behaviour analysis, multiple target tracking, and so on.

These types of intelligent functions require a series of front-end surveillance cameras, back-end servers and other products which support deep learning algorithms. In small scale applications, front-end cameras can directly operate structured human and vehicle feature extraction, and tens of thousands of human facial images can be stored within the front-end devices to implement direct facial comparison, so as to reduce costs of communicating with a server. In large scale applications, front-end cameras can work with back-end servers. Specifically, the structured video task is handled by front-end devices, reducing the workload for back-end devices; matching and searching efficiency of back-end servers improve as well.

Hikvision new products with Deep Learning

This year, Hikvision will soon introduce a series of products with deep learning technology, such as the DeepInview Series cameras which can accurately detect, recognise, and analyse human, vehicle, and object features and behaviour, and can be widely used in indoor and outdoor scenarios. Another of products worth mentioning is Hikvision’s DeepInmind Series of NVRs which incorporate advanced deep learning algorithms and imitate human thoughts and memory. The DeepInmind products feature an innovative NVR+GPU mode, retaining the advantages of traditional NVRs and additional structured video analysis functions, which together greatly improve the value of video.

Deep learning is the next level of AI development. It is beyond machine learning where supervised classification of features and patterns are set into algorithms. Deep learning incorporates unsupervised or “self-learning” principles. Hikvision is developing this concept in its own analytics algorithms. Enhanced accuracy is the result of multi-layer learning and extensive data collection. Application of this algorithm into face recognition, vehicle recognition, human recognition, and other platforms will significantly advance the performance of analytics.

Download PDF version

In case you missed it

Smart home access control growth and the future of door security
Smart home access control growth and the future of door security

There’s growing noise around smart homes and smarter security. You’ve probably heard it. But there is a place where access control and more have been smart for decades: the workplace. Home automation and IoT are still playing catch-up with the commercial sector. A new insights report from ASSA ABLOY and IFSEC Global — “The Smart Door Locks Report 2018” — measures just how fast consumer smart technology is running. According to a survey conducted for the report, 61% of households now claim to own at least one smart home device or system. Energy monitors, home CCTV cameras, intruder alarms and smart door locks are the most popular, according to the report. All these functions, of course, have been available to businesses for years.61% of households now claim to own at least one smart home device or system Educating the smart home consumer Paradoxically, report data also questions how much consumers really know about their smarter home. A surprising 42% of those surveyed, for example, were unaware they could control a smart door lock from their phone. In fact, many leading smart door lock models offer this feature, delivered by Wi-Fi or Bluetooth and an app. Despite a wealth of features offered by the latest smart door locks — remote and location-based locking/unlocking; voice activation; timed access; emailed entry alerts; and integration with smart camera and lighting systems — most people still have limited knowledge of their capabilities.  Smart technology is increasingly becoming the new norm in terms of home security  Only 14% of survey respondents described themselves as “very familiar” with what a smart lock can do. Even though most of them probably use smart access control solutions at their workplace. Secure homes through smart technology Monitoring and security are not the only drivers for smart home adoption. We humans also love convenience, and modern living presents us with problems that smart home technology can solve. Ironically, given the report’s findings, it takes a smartphone to really unlock the convenient possibilities of smarter living. The device that’s “always to hand” is central to the newest generation of smart door locks.A smart door lock is a convenient way for a landlord or agency to offer round-the-clock check-in and check-out If homeowners wish to remotely manage property access for friends and family, many smart door locks oblige. You let in guests remotely, send them a virtual digital key, or provide a temporary or single-use PIN to unlock the door. It is just as easy to revoke a digital key, if you don’t want its owner to come around anymore. This is a significant improvement over sharing physical keys — or hiding one under the doormat. We cannot be totally sure where a metal key ends up and have no way to track or cancel it once it’s “out in the wild”. Commercial access control offers such functionality as standard, of course.  In addition, smart door locks offer more than just stand-alone operation and clever functions. In a domestic setting, magic happens when locks work in harmony with a home automation system, connected by protocols like Z-Wave, ZigBee or Wi-Fi. "Smart" security on the move  The smartphone is becoming a remote control for managing a connected life beyond just home (and even workplace) security. According to Accenture, the parcel delivery services market will grow by $343 billion by 2020. Just like home security, convenience is a major driver of change. Homeowners can send guests a virtual digital key to their phones, or provide a temporary or single-use PIN to unlock the door A recent PostNord pilot in Sweden aimed to remove the inconvenience of waiting home for a postal delivery. Selected customers of some major Scandinavian e-retailers could choose to have parcels delivered inside their front door, if it was equipped with a Yale smart door lock.  Home delivery is among potential smart services covered in “The Smart Door Locks Report 2018 ”. When asked whether the ability to receive parcels securely in a porch or lobby would make them more likely to invest in a smart door lock, 79% said it would.It is easy to revoke a digital key, if you don’t want its owner to come around anymore Holiday rentals and smart home tech ASSA ABLOY research published in 2017 forecasts continued growth in the European holiday rentals sector (at 5.8% CAGR). Smart door locks are also making an impact here, at both ends of the market: for service providers — agents and homeowners — and for travellers. A smart door lock is a convenient way for a landlord or agency to offer round-the-clock check-in and check-out, without creating extra work or staff costs. Both Intersoft, in Croatia, and Hoomvip in Spain have built holiday rentals management systems around an app and the ENTR® smart door lock. Agents issue, revoke, track and manage virtual keys for all their guests, saving everyone time and hassle. Travellers use their phones and an app to unlock their apartment. For these visitors the smartphone is already an essential travel accessory. It is a boarding pass, a credit card, a travel guide, and a postcard home... why not a door key, too? And if this key is backed by a trusted home security brand — and a company with vast experience in the mature market for commercial “smart” security — better still.

Bosch startup SAST addresses need for evolved solutions in security industry
Bosch startup SAST addresses need for evolved solutions in security industry

Security and Safety Things GmbH (SAST) is a new company that has announced its vision for an Internet of Things (IoT) platform for the next generation of security cameras. The Bosch startup plans to build a global ecosystem for the development of innovative security camera applications. Based on the Android Open Source Project (AOSP), SAST provides libraries, an API framework, and codecs for developers to work with. The SAST App Store will allow developers to build and market new applications, similar to today’s app stores for smartphone applications. We presented some questions to Nikolas Mangold-Takao, VP Product Management and Marketing, about the new venture, and here are his responses: Q: Why a new company now? What technology innovations have made this a good time to launch this company? The time is right to bring market needs and technological innovations together on one platform"Mangold-Takao: From a technical perspective we see two main drivers: increasing computing power at the edge and increasing internet connectivity, which will enable devices to directly communicate with each other and bring new technologies such as artificial intelligence also to the security and safety industry. At the same time, we see that this industry and its users are hungry for more innovative solutions – addressing new security needs while at the same leveraging the possibility to improve business operations for specific verticals, e.g. retail and transportation. The time is right to bring market needs and technological innovations together on one platform for this industry. Q: Why does SAST need to be a separate entity from Bosch? Mangold-Takao: SAST is setup as a wholly owned subsidiary of the Bosch Group. We wanted to make sure that SAST is able to underline its role as an industry standard platform across multiple players. SAST is open to get additional investors and is being setup as a startup in its own offices in Munich to foster the environment where speed and innovation can more easily take place. Having said that, several entities of the Bosch Group are very interesting partners for SAST. The SAST App Store will allow developers to build and market new applications, similar to today’s app stores for smartphone applications Q: Please explain your "value proposition" to the industry. Mangold-Takao: We will bring new innovations and possibilities to the security and safety industry by providing an open, secure and standardised Operating System for video security cameras, to also address pressing issues such as cyber security and data privacy concerns. Devices that run then with the SAST operating system will work with an application marketplace provided and operated by SAST. Integrators and users can then use these apps from this marketplace to deploy additional functionality on these devices. With our platform we will be able to build up a community of app developers, including the ones not yet developing for this industry who have expertise in computer vision and artificial intelligence. Q: It seems what you are doing has parallels with the Apple and Android "app" stores. How is your approach the same (and how is it different) than those approaches? We are setting up SAST as a user-centric company and involve selected users very early on in the process"Mangold-Takao: The approach is similar in the way that we plan to generate revenue by operating the application marketplace and thus participate in the app revenue. The difference is that there is much more needed than apps and cameras to create a complete working solution addressing a user problem in this industry – we need to make sure that our own platform as well as the new applications being created will work as a part of an end-to-end solution. Q: "Critical mass" and wide industry participation seem to be requirements for your success. How will you achieve those goals? Will you involve integrators, consultants, or other parties in addition to manufacturers (to drive awareness)? How? Mangold-Takao: SAST is in close exchange with device manufacturers, integrators and consultants, as well as application developers and large end-users at the moment to ensure that we are building the right platform and ecosystem for this industry. We are setting up SAST as a user-centric company and involve selected users very early on in the process. We will run dedicated programs and hackathons to attract app developers, already active and new to our industry. We will also run selected pilots with end-users throughout 2019 to ensure we have all partners involved early on. SAST sees the industry is hungry for more innovative solutions – with the retail vertical market a target for these solutions Q: What timeline do you foresee in terms of implementing these initiatives? Mangold-Takao: While we start with first app development programs and plan our first pilots already for this year, we are planning our commercial launch for end of 2019. Q: How does your new company relate to the new Open Security & Safety Alliance (OSSA)? Mangold-Takao: The Open Security and Safety Alliance has been working very closely with SAST over the past year, defining some important concepts and elements required. One of the most important elements is an open and standardised Operating System, specific to this industry, which will then bring forward new innovative technologies and solutions. SAST is actively working on this Operating System, based on Android Open Source Project (ASOP), but is evolved and hardened with industry-specific features. Q: What's the biggest thing you want the security industry to understand about SAST? What is your "message" to the industry? Mangold-Takao: Our message is simple: let’s build better security and safety systems – together! But for real, innovating an industry is a joint effort, we can only bring new innovation to this industry with partners who share our vision and are excited about new technology. At the same time, we strongly believe that our platform allows every partner to bring forward what they do best but also invite new partners to our industry.

What is the value of remotely monitoring a system's health and operation?
What is the value of remotely monitoring a system's health and operation?

When is it too late to learn that a video camera isn’t working properly? As any security professional will tell you, it’s too late when you find that the system has failed to capture critical video. And yet, for many years, system administrators “didn’t know what they didn’t know.” And when they found out, it was too late, and the system failed to perform as intended. Fortunately, in today’s technology-driven networked environment, monitoring a system’s health is much easier, and a variety of systems can be deployed to ensure the integrity of a system’s operation. We asked this week’s Expert Panel Roundtable: How can remote monitoring of a security system’s health and operation impact integrators and end users?