The “2023 Tech Trends Maturity and Suitability Index Survey,” sponsored by ZKTeco, garnered responses from 633 channel players offering validated insights into the technologies significantly impacting the video surveillance and access control industry. This year’s survey includes a closer look at biometrics, with insights into the industry and emerging trends from key market players and a comprehensive analysis of the survey results.
Ongoing and Emerging Trends
Mature and suitable: Radar and LiDAR (2), edge processing/storage (3), unmanned security platforms (4), artificial intelligence (5) and 5G wireless communications (6); Mature but not as suitable: Li-Fi (1)
5G ranks high in the survey and for good reasons too; the technology with its fast speed and low latency has been a key enabler in AIoT and is now increasingly seen in video surveillance with more cameras being 5G-enabled. Also ranking high are edge processing/storage, unmanned security platforms (robots and drones) and radar/LiDAR, the latter of which are companion detection technologies to video surveillance, being able to detect objects in inclement conditions. LiFi or light fidelity is becoming mature yet is still deemed unsuitable as a transmission solution.
Access Control
Mature and suitable: Touchless biometrics (7) and building/property management integration (8); Mature but not as suitable: Unimodal biometrics (1), touch biometrics (2), wireless locks (3), cloud access control (4) and digital twins (5); Not as mature nor suitable: Multimodal biometrics (6)
Touchless biometrics such as face and palm recognition are gaining ground post-pandemic while building management integration suits the current green and sustainability trends well. Unimodal biometrics and touch biometrics, while technologically mature, rank lower in suitability as organizations attach stronger importance to the security, safety and health of their workers. Multimodal biometrics, on the other hand, is moving up along the suitability index due to the extra layer of security it offers.
Video Surveillance
Mature and suitable: Mobile apps for video surveillance (7), low light color cameras (6), 4K resolution (5), cloud video surveillance (4) and bi-spectral camera (3); Not so mature nor suitable: 8K resolution (1) and tri-spectral camera (2)
Lowlight color cameras offer good color-capturing capabilities in lowlight conditions and continue to gain ground in video surveillance thanks to more advanced lenses and image processors available. 4K allows users to zoom in and still see clear images, making it useful in large-scene environments such as city surveillance. Bi-spectral cameras whereby visible light and thermal camera units are combined into one device are beneficial in industrial settings. By comparison, 8K resolution and tri-special cameras (with UV combined) are not as suitable, at least at this point in video surveillance.
2023 Video Surveillance Tech Trend Survey: AI Dominance Continues
In this year’s asmag.com video surveillance technology survey, sponsored by ZKTeco, lowlight color cameras, 4K and bi-spectral solutions all ranked high in both market suitability and technical maturity, while AI continues to gain traction as a solution that fulfils users’ security and operational needs. This article takes a closer look at the results.
AI
AI, in particular AI predictive analytics and anomaly detection, ranks very high on the suitability index. This is understandable, as more and more users rely on AI to achieve better security and efficiency.
“2023 saw the growing influence of artificial intelligence (AI) as data-driven video technology, the combination of AI and video technology. In data-driven video technology, instead of people watching the video, the software watches the video,” said Rahul Yadav, CTO of Milestone Systems. “Data-driven video technology is disrupting the security industry, but it’s not removing people from the solution; it’s moving people to the center of the solution. Software is now the tool at the periphery doing the watching, identifying objects, recognizing patterns, trends, and correlations, and using these to create insights and actionable intelligence.”
“In the video surveillance market, manufacturers are continuing to focus on AI-based analytics. i-PRO’s Scene Change Detection is an example of how analytics are evolving beyond recognizing people, vehicles and attributes,” said Adam Lowenstein, Director of Product Management at i-PRO Americas. “At the same time, we have heard more issues and concerns about how analytics are being used. Privacy concerns are being highlighted more often due to AI being frequently in the news, along with misconceptions about the difference between AI and facial recognition. As a result, we’ve had more enquiries about our edge-based AI Privacy Guard feature that applies a mosaic to the face or entire figure of a person.”
“In the coming years, we will see innovative surveillance solutions come to market that are more cost-effective through the cloud and smarter through AI. In particular, we can expect to see Edge AI cameras that capture, store, analyze, and connect to the cloud from the edge of a single device, instead of having to build many separate systems and infrastructure such as video devices, storage servers, and AI analysis servers,” said Choong Hoon Ha, Chief Sales and Marketing Officer at Hanwha Vision.
As far as technical maturity is concerned, AI predictive analytics and anomaly detection ranks in the middle. But that’s sure to change once better chips and algorithms become available to make AI even more mature and reliable.
Lowlight Color Cameras
Lowlight color cameras rank highly in both suitability and maturity. Indeed, thanks to better lenses, sensors and chips, coloring capturing in extremely low Lux conditions has become increasingly feasible, and for users who attach great importance to nighttime surveillance, this has proven beneficial for them.
“Low-light color security cameras are commonly used in critical infrastructure facilities such as power plants, water treatment facilities, airports, and other major transportation hubs often as an alternative or in combination with thermal imaging cameras. Low-light cameras help ensure security and integrity of vital sites, even during nighttime or low-light scenarios,” said Jamie Barnfield, senior sales director, IDIS Europe. “They are also useful to monitor entrances, exits, and interior spaces that experience challenging light conditions, allowing for accurate identification of activities in real-time and deliver crisp, clear recorded footage.”
“Since i-PRO has been ahead of the curve with low-light sensors delivering excellent color reproduction for some time, it is not what we would describe as a new trend. However, we have worked hard to educate the market on the benefits of capturing color in low light,” Lowenstein said. “Beyond the obvious benefits for humans to better identify a person or vehicle, AI-based object recognition also functions much better when it has color information to work with.”
UHD
4K ranks highest for maturity and at the high end for suitability. Thanks to better sensors and codec technologies, 4K or 8MP security cameras can capture UHD images while keeping bandwidth consumption and storage at low. This makes 4K especially ideal for scenarios where the user needs to, for example, establish the pattern of a person’s clothes or identify the plate number of a vehicle far away. But does it mean the higher the resolution, the better? Not necessarily. Consider 8K, which ranks low in both suitability and maturity. Further, UHD may be more expensive, and they may work less effectively in lowlight conditions.
“Since the 4K market is still growing compared to Full HD installations, 8K use cases represent a tiny minority for most customers today. Many users have found that having a combination of a multi-sensor camera which covers 360 degrees in HD or 4K coupled with an auto-tracking PTZ camera that can optically zoom in much closer, is more practical, flexible and affordable. 8K also impacts storage and bandwidth considerably compared to 4K cameras. It also takes extremely good lighting conditions to benefit from 33 megapixels of resolution in most security installations,” Lowenstein said.
“It is still true that lower resolutions are the more cost-effective solution. Moreover, they are still absolutely sufficient for many applications. Especially in the low-light range, the lower resolutions are currently even better, since larger pixel sizes capture more light. This is why MOBOTIX also offers a 4MP low-light sensor for the MOBOTIX 7 series,” said Thomas Lausten, CEO of MOBOTIX. “But we are convinced that the more demanding an application is to be, the higher resolutions will be required.”
Bi-spectral Cameras
Also ranking high on the chart is bi-spectral cameras where a visible camera unit and a thermal camera unit are combined into one. This is especially ideal for factory and industrial settings where detection and verification of smoke, heat and fire points are critical. Now there are also tri-spectral cameras with a third, UV component added. But these rank lower in the survey.
“There is already interest in some fields outside of the security industry such as environmental monitoring and agriculture as tri-spectral scans capture images in different lighting conditions and provide valuable insight and detail that is not visible to the naked eye. Yet for the time being, multispectral and bi-spectral cameras are most in demand to improve the accuracy of many AI-powered analytics applications, that are unable to hit the ‘sweet spot’ of up to 98 percent accuracy for people counting in retail,” Barnfield said.
Also, it should be noted that whether visible, bi-spectral or tri-spectral, these are just means to an end. “For the time being, it is not important whether the image is optical, IR or UV-based. That is a decision defined by the task at hand and the environment. What is decisive is the comprehensive intelligence with which the system is supported. This was also the prevailing trend in 2023: artificial intelligence-based software that, in combination with a camera, correctly detects situations that precede unwanted events to ultimately prevent them, but also to optimize processes,” Lausten said.
Other Items on the Survey
Cloud video surveillance ranks high, allowing users to opt out of investing in NVRs and servers. SSD in storage is also gaining prominence, presenting a faster and more reliable storage solution. Finally, solar-powered video surveillance is gaining ground in market suitability, providing a viable solution in remote areas where power is not readily available.
2023 Access Control Trend Survey: Touchless, Multimodal Biometrics Gain Ground
This year’s Security 50 access control technology survey sees building management integration and touchless biometrics both ranking high in terms of technical maturity and market suitability. Meanwhile, multimodal biometrics, where users are authenticated by more than one biometric modality, are gaining ground. This article takes a closer look.
Building Management Integration
Building management integrated with access control ranks the highest in suitability. In terms of maturity, it clinched a high spot as well. Indeed, more and more users are turning to building management-access control integration due to its various benefits, including one-stop management and energy savings.
“The modern building environment is a complex network of numerous systems, including air conditioning, water drainage, plumbing, electricity, fire alarms, security systems, and home or office automation systems. Managing each of these systems independently can be a daunting task, often leading to inefficiencies and an increased likelihood of errors. More and more building owners and operators are recognizing the value of a unified platform that brings together the management of all these systems,” said Raymond So, Head of Marketing at ZKTeco.
He added: “In a well-integrated building management system, there is a high degree of interaction between different systems and devices. For instance, an access control system can be integrated with the lighting and HVAC systems to automatically adjust the environment based on occupancy. This level of interaction enhances building security, increases energy efficiency, and improves the overall user experience by creating a more intuitive and responsive environment.”
Touchless vs. touch biometrics
For biometrics, the survey shows touchless biometrics such as face and iris recognition rank high in both suitability and maturity. Touch biometrics, meanwhile, ranks the highest in maturity. This is understandable, as touch biometrics, mainly fingerprints, have been in use for decades. Yet, the value of touchless biometrics is also being realized, especially in the post-pandemic era.
“Touch biometrics have been around for decades and are widely deployed for access control applications. So, yes, we agree they are more mature in terms of their acceptance and maturity in the marketplace,” said Brian DeGonia, Director for Biometric Solution Enablement, Extended Access Technologies at HID. “Touchless biometrics however are gaining ground quickly due to the frictionless, ease of use and the ability to operate without contacting a surface that may be of health concern. The performance of touchless matching algorithms is improving quickly and drastically to have the same characteristics as their touch counterparts.”
“The growing prominence of touchless biometrics is shaped by factors such as hygiene, user-friendliness, speed, cultural considerations, and market trends. Despite this, touch-based biometrics continue to hold their ground owing to their maturity and familiarity among users,” So said. “At ZKTeco, we are committed to advancing both touch and touchless biometric technologies to cater to our diverse global clientele. Our focus remains on delivering cutting-edge, reliable, and user-friendly biometric solutions that align with evolving market needs and trends.”
“AI-based identity solutions and touchless biometrics have gained deep adoption in the access control realm. This is a result of the outcome of a sustainable solution during the pandemic outcry for a safe and secure access control solution, requiring a hands-free contactless proposition. The advent of AI and the steep innovation and improvement over the recent years has also accelerated and pushed the adoption curve towards a new behaviour and trust for contactless biometrics for access management and identity management,” said Alex Tan, Regional Sales Head for ASEAN at IDEMIA. “The requirement for hard credentials or digital cards will get softer by each year as ESG initiatives and governance come center stage for every enterprise – be it at the local or governmental level. This is also driven by the fact that biometrics are virtually perpetual requiring little to no credentialing after support.”
Multimodal Biometrics Pick Up Steam
The survey further finds unimodal biometrics ranking higher on maturity, while multimodal biometrics are not as mature but are gaining ground in suitability. Indeed, multimodal biometrics are seeing increased adoption due to certain benefits.
“These include higher accuracy, where multimodal biometrics can reduce the error rates and increase the confidence levels of the system by combining the strengths of different modalities and compensating for the weaknesses of each one; higher security, where multimodal biometrics can prevent spoofing attacks and enhance the privacy of the users by requiring more than one biometric trait to grant access; and higher user convenience, where multimodal biometrics can offer more flexibility and choice to the users by allowing them to select the most suitable and comfortable modality for a given situation,” said Hanchul Kim, CEO of Suprema.
However, multimodal biometrics also have some challenges and limitations that make them less mature than unimodal biometrics. These include higher cost and complexity, and less stable infrastructure. Yet these can be overcome in time.
“Multimodal biometrics have a great potential to overcome their current limitations and become more mature and widely used in the future as the technology evolves and the market demand changes over time,” Kim said.
“Multimodal biometrics are increasing in use for identification. For authentication, unimodal works quite well and is very mature. But for large-scale identification use cases or in applications that require a higher level of security and assurance, multimodal biometrics can be used to disambiguate where an unimodal biometric may match more than one person. By using more than one biometric factor, you get better assurance that you’ve identified the correct person,” DeGonia said.
However, this is not to negate the popularity of unimodal solutions which some said would stick around for a while. “This is in line with human behavioural mechanics where people are simply more geared towards doing simpler or lesser gestures to attain the same outcome. Take for instance making a phone call, one with the early 1900s rotary dial gesture versus the late 1960’s push-button gesture. Even despite saving only a few seconds in the process, convenience and speed is the driver that brought about the demise of rotary dial phones. As convenience and accuracy take center stage, touchless unimodal authentication like facial recognition or fingerprint technology will make up the vast majority of authentication modalities. It is hard to beat the convenience of not bringing anything but yourself (something that you are) instead of having to hold or wear something (something that you have or carry),” Tan said.
He added: “In the past, only sensitive areas/doors will require multimodal technology to ensure effective security posture at that area. The same is true for the field of IAM like making important transactions like payment. Thus, multimodal in its true use case is meant for important transactions; otherwise, the inaccuracy may happen in certain unimodal technologies. Hence, the value of any biometrics is to get its 1:N function to work accurately with speed and inclusion. Just look at how we unlock our phone, do we use multimodal to unlock it? No. multimodal is only to heighten the security stance for special environment/use case.”
Dual Biometrics
At this time, multimodal biometrics are mostly dual-modal, namely using two modalities to authenticate. According to our survey, face-fingerprint and face-palm rank higher in both suitability and maturity than palm-fingerprint.
Our interviewees mostly agreed with the results. “When thinking about duo-modal biometrics, you often look to use different, distinct body features. A hand consists of both the fingers and the palm. As a result, you commonly see face-fingerprint implementations since these technologies are readily available in the market and in form factors that can be easily integrated,” DeGonia said.
“Face fingerprint is one of the most common and popular duo-modal biometric technologies, as it combines two of the most widely used and reliable biometric modalities. Face-palm is a relatively new duo-modal biometric technology, but it has gained a lot of attention and interest in recent years. It combines face recognition with palm print or palm vein recognition, which are both contactless and robust modalities. It has been applied in some emerging applications, such as health care, banking, and education. Face-palm offers high performance, privacy, hygiene, and flexibility,” Kim said.
He added: “Face-fingerprint and face-palm are both suitable for applications that require high levels of security and convenience. They can provide fast and accurate verification or identification with minimal user interaction and effort. They can also prevent spoofing attacks and protect user privacy by requiring more than one biometric trait to grant access.
According to Kim, palm fingerprint is a less common and less mature duo-modal biometric technology. “It combines two contact-based modalities that require different sensors and capture methods. It has been used in some niche applications, such as forensic investigation, and criminal identification,” he said. “Palm-fingerprint is less suitable for applications that require high levels of security and convenience. It can provide high uniqueness and diversity of biometric features, but it also requires more user interaction and effort to capture both modalities. It can also be affected by environmental factors, such as dirt, moisture and temperature that can degrade the quality of the biometric images.”
