Hikvision, an IoT solution provider with video as its core competency, announced a brand-new addition to its DeepinView camera line: the Dedicated Subseries. This unprecedented new addition loads a batch of AI-powered deep learning algorithms into each unit, boasting stunning performance and cost-effective pricing.

Over the last few years, artificial intelligence (AI) has been applied in many ways in security markets. As technology advances, AI chipset performance has improved to enable massive computing power using various algorithms and contributing to multi-intelligence functionality and higher accuracy. The new Dedicated DeepinView Cameras are an example of these advances, incorporating several AI-powered deep learning algorithms in one unit. What’s more, these algorithms can be switched essentially putting 5 or 6 unique cameras in one housing.

“Embedding switchable algorithms is a significant step for Hikvision to take in its AI product development. In a world of ever-changing technologies and functionalities, this approach creates great value for end users to try new technologies to ensure security, as well as to implement business intelligence and other applications,” says Frank Zhang, President of the International Product and Solution Center at Hikvision. “The benefits of our new offerings are numerous including reduced costs, improved efficiency, and speedy and effective incident response.”

Switchable algorithms

The Dedicated DeepinView cameras combine two product categories – the first is vehicle analysis where cameras combine automatic number plate recognition (ANPR) with vehicle attribute recognition. Attributes include the vehicle’s make, color, and direction of movement. Typical uses include installation at checkpoints of city streets and at entrances & exits of buildings or industrial parks.

Models in the second category boast six switchable deep learning algorithms in one camera housing, including facial recognition, face counting, hard hat detection, perimeter protection, queue management, and multiple-target-type detection (detecting multiple targets and multiple types of targets at once). Accordingly, users can simply enable an algorithm manually for dedicated use, then later switch the algorithm as needed.

Here is one example: hard hat detection. This algorithm can be used on construction sites to ensure safety and compliance. Specially-equipped DeepinView cameras can precisely distinguish a worker on the site wearing a hard hat from those without, and automatically deliver alerts when the hard hat violation is detected.

Another example: in a retail setting, a face-counting function can be enabled to precisely count customers entering and leaving the store. Repeat customers and store staff can be automatically excluded in the process, helping store managers count new customers with precision.

Flexibility among algorithms enables users to also switch among:

Perimeter protection – to monitor outdoor areas needing security and deliver accurate alarms upon intrusions.

Facial recognition – to grant authorized access to restricted areas in various organizations, such as school laboratories, archive rooms, and hospital pharmacies.

Queue management – to better understand customer wait times, optimize staff levels, and enhance customer experience.

HD clarity, day and night

Equipped with Hikvision’s DarkFighter and LightFighter technologies, these cameras capture vivid and color images in extremely low-light environments or in scenes with strong backlighting where color and brightness balance is extremely difficult. Smooth Streaming mode further ensures a high-quality live feed.

The Dedicated DeepinView Cameras are available in 2, 4, 8, and 12 MP resolutions for customers to choose from.

More practical and deployable features

Furthermore, metadata is supported to allow third-party platforms to receive data from Hikvision cameras for real-time video analysis or recorded into footage archives to enable rapid searching forensic evidence.

Finally, these camera models also offer Vibration Detection for outdoor use, which detects and notifies users of vandalism.

The post Hikvision Introduces Dedicated Series In Its DeepinView Camera Line appeared first on Artificial Intelligence.

As we enter the new decade, artificial intelligence is both one of the most innovative and the most polarizing emerging technologies out there. It is the electricity powering the fourth industrial revolution, progressing at such a fast pace that we may not understand its full potential for decades to come. Over the last 30 years, AI has turned from something extremely specialized to a program in and of itself. It is being embedded into technology we use every day, like our mobile devices, and also into new, broader initiatives like smart cities.

As with many revolutionary innovations, there is a race to be the fastest to tap into the benefits of AI. China, for example, is racing to showcase the best uses of AI technology, deeming themselves the “AI Utopia.” According to the National Security Commission on Artificial Intelligence, China’s targets are ambitious. In the next ten years China wants country-wide adoption of AI in manufacturing, health care, education, agriculture and defense, and is setting its sights to be the global leader, against the US, by 2030.

I know you

In 2020, we will see US governments shift the conversation from who implements AI fastest to how we can implement most responsibly. While China is already using AI to measure students’ brain waves with IoT sensors during class to help teachers provide more customizable content to achieve better retention and results, it’s likely that the U.S. government will focus heavily in the coming year on privacy regulations to ensure AI use cases like this are fully vetted before being allowed.

Federal regulations on privacy when it comes to the use of AI will take center stage in 2020. We’ve already seen the beginnings of this with two instances of the U.S. government taking action to prevent AI overstepping in states California and Massachusetts. This past May, the San Francisco Board of Supervisors banned the use of facial recognition technology by police and all other municipal agencies under the Stop Secret Surveillance Ordinance. Similarly, Sommerville, Mass. voted on a facial recognition ban for police investigations and municipal surveillance programs in July, 2019.

While it may seem simple enough to ban any potentially invasive uses of AI, facial recognition algorithms also have the capability to aid law enforcement officials in investigations, enabling police departments to scan massive amounts of footage in attempt to find a person of interest, for instance. 2020 will be about finding the balance between reaping the benefits of AI and making sure it doesn’t go too far.

Be prepared for unparalleled AI-powered innovations to come over the next year, but also be prepared for privacy, security and AI ethics to be at the forefront of the conversation. As regulations continue to be made at a city, state or federal level, they will shape the extent to which we can use AI, and how we will approach the “AI race” in this decade.

More in AI & Machine Learning

The post AI and facial recognition in 2020: where’s the line? appeared first on Artificial Intelligence.

The global videos surveillance market is expected to post a compound annual growth rate of close to 11% during the period 2018-2022 according to Technavio. The potential benefits of leveraging artificial intelligence (AI) in the physical security industry have pros and cons on both sides, but the debate over the ethical ways to leverage AI and surveillance continues as more and more surveillance systems are getting the brains to match what they see.

AI startups like Boulder AI, which offers a vision-as-a-service and IC Realtime, which lets you search and analyze your video feeds from CCTV system; are gaining traction. Alongside the Chinese facial recognition startups like Megvii’s Face++ with $600 million in private equity; SenseTime with $62o million from a series C; and, Yitu Technology with $300 million from a series C, the potential uses of facial recognition technology are well funded.

Umbo Computer Vision is a Taiwanese startup with $10 million for its AI-powered video security system that can understand human behaviour. 

“In the future, when we are walking on the street and see an AI-powered camera, we will feel safe,” said Guan.

Guan believes that AI is the key to transform the security industry from evidence collection to prevention. “Machine learning and AI to understand human behavior unlocks the passage to real-time situation awareness of public safety and asset protection, like making sure no one can scale the wall to come into an elementary school, break into cars or facilities.”

“In three to five years from today, AI will be in all conversations about physical security systems,” he adds.

Mike Ellenbogen, is CEO and c0founder of Evolv Technology, a threat detection system that uses facial recognition to screen people’s faces. Evolv has $30 million in funding from General Catalyst, Lux Capital, Gates Ventures and Data Collective.

“The number of mass shootings in the U.S. continues to rise – with three of the deadliest shootings occurring within the last year. Adversaries are getting more advanced and targeting large open spaces where mass groups gather. Society needs a new approach to security, and AI and machine learning (ML) are playing a pivotal role in transforming the physical security landscape,” said Ellenbogen.

Ellenbogen says that facial recognition is also helping to improve security by identifying individuals entering a venue or business.

“The technology can verify a person from a digital image or video database to determine if they should enter,” adds Ellenbogen. “Whether it is preventing someone on a watch list or only allowing known users in – the algorithm is helping security professionals proactively monitor and prepare for threats. It’s important to note, however, that facial recognition isn’t enough. A larger security plan needs to integrate identification tools with threat detection technology to truly prevent unauthorized entry.”

Taking that one step further, AI’s adoption in security is still evolving, and for now, it can’t work alone.

“The technology can analyze data and identify patterns but can’t always determine if each pattern is an immediate danger. This requires insightful thinking that only the human brain is capable of. In the end, it will be the combined power of AI and human intelligence that will allow us to combat today’s constantly changing threat landscape,” said Ellenbogen.

“With biometrics and AI embedded into security technology, we’re empowering our customers to take a proactive approach to security and have the capabilities to expand security perimeters beyond the front door,” said Ellenbogen.  “One of the key differentiators AI offers [..] is the ability for a system to learn and one common application of ML we see in physical security is object recognition in which the system is taught to identify an object as a threat based on certain characteristics – such as the signature of a gun, knife or bomb.”

Ellenbogen says that today’s advanced ML does a better job compared to machine vision approaches developed just a few years ago.

“The result is that we have systems that can do more of the monitoring automatically, and then quickly alert guards of a potential threat. Automatic monitoring of these environments reduces the need for full-body pat downs, creating a less invasive and time-consuming security environment that requires less labor,” said Ellenbogen. “In larger environments, sensors can determine if someone’s bag might contain a threat and then track that visitor down.”

“We strongly believe that facial recognition is not meant to implicate people and prevent them from entering an event or building, it’s simply used to alert a guard of a potential match, so a guard can better evaluate the situation,” said Ellenbogen. “The combination of AI/facial recognition and human IQ is what helps a human being make a sound judgment call about whether to evaluate a person further or adjust the system sensitivity. It’s incumbent upon society, to become more comfortable with opting in and consenting to the use of relevant identity data for the sake of safety,” said Ellenbogen.

Guan says that the greatest beauty of the human race is that we are able to adapt quickly.

“Some draw connections to the rise of AI as the industrial revolution. There were concerns back then about machines taking away jobs and ultimately, people learned to coexist with machines and use them to push the human race forward,” said Guan. “Today, people have similar concerns to the changes that AI could bring, yet one day AI will not be something we talk about as it will become an integrated part of our life. The security industry shares the same faith.”

Ellenbogen points out that in the current threat landscape, we will need integrated security solutions that allow people to walk into a familiar place without stopping, whether it be at a stadium, an office building or a train station.

“While there continue to be discussions around the right or wrong way to leverage advanced technology, such as facial recognition, if it can prevent one mass casualty or mass shooting event from happening, that’s a trade we should be willing to make,” said Ellenbogen.

The post Making Facial Recognition Smarter With Artificial Intelligence appeared first on Artificial Intelligence.