SecurityWorldMarket

03.12.2020

Intelligent Security Systems - Part 4 of 6

The state of AI in the surveillance sector

Artificial intelligence (AI) is a branch of computer science that studies and develops methods that allow computers to simulate intelligent behaviour.

Artificial intelligence (AI) is a branch of computer science that studies and develops methods that allow computers to simulate intelligent behaviour.

Everybody is talking about artificial Intelligence, AI, these days. But what is AI? How do we define it in the security industry?
Detektor decided to interview Mats Thulin, Director Core Technologies, at Axis Communications regarding AI and its potential in surveillance applications.

Artificial intelligence (AI) is a branch of computer science that studies and develops methods that allow computers to simulate intelligent behaviour.

“In general terms AI is a very broad concept, but in the specific context of video analytics the principal focus is to increase operational efficiency and add value by automatically processing and analysing video streams,” states Mats Thulin.

“In this context a subcategory of AI, machine learning (ML), is more specifically relevant. As its name suggests, machine learning allows computers to improve algorithms through ‘learning’ based on real-world examples. The improved algorithms are then used to analyse images or video sequences to generate alarms, metadata or other information.”

Deep learning

Mats Thulin, Director Core Technologies, at Axis Communications.
Mats Thulin, Director Core Technologies, at Axis Communications.

Furthermore Thulin claims that attention has the last few years turned to a subcategory of ML, deep learning (DL), which describes algorithms based on simulated neural networks.

“The idea for this type of algorithm was inspired by the human vision system, hence its name – neural networks. In DL networks, layers of operations are arranged in a hierarchy of complex and abstract layers, each layer using information from the previous one to draw its final conclusion,” he says.

Mats Thulin emphasises that DL models enable more complex analytical algorithms and generally achieve greater precision than traditional ones.

“In video surveillance systems they are used primarily in the detection, classification and recognition of different types of objects. However, one drawback of DL algorithms is that they require more computational power and more mathematical operations in comparison to traditional algorithms.”

Demand for lots of data

ML and DL require huge amounts of relevant input data for training to achieve good quality results. If enough relevant data – and computing power – is available for training, ML- and DL-based methods can efficiently process it to achieve algorithms with higher precision. The computer can analyze thousands of images to find details that characterize specific objects in different scenarios.

“If the data and their descriptions are of high quality, an application based on DL is able to achieve even greater accuracy. But availability of high-quality data can be a challenge,” claims Mats Thulin.

“Perhaps countering the general perception of AI, today’s technologies still lack awareness or what might be referred to as general intelligence. In applications where the technology is used, it focuses on very specific problems in limited domains. For example, for a voice application such as Siri or Alexa to accurately answer our questions, we need to ask very specific and explicit questions. Otherwise, we will get a completely incomprehensible answer. Similarly, in surveillance systems, a poorly defined use case without proper delimitations may result in applications achieving low accuracy.”

Based on the current limitations in accuracy Mats Thulin emphasize that we must be cautious on how and where to use these technologies.

“The technology today improves efficiency but the actual decision making in a surveillance scenario must still lie with the security guard or the operator. We must keep a ‘human in the loop’, he says.

The surveillance in focus

As any new technology matures beyond the initial ‘hype’, weaknesses and limitations of the technology will become clear and only in the areas where the technology provides real value we will see growth, according to Mats Thulin.

“In surveillance it is important to start with the use-case: what problem are you trying to solve, or which effect are you looking to achieve? Based on a good understanding of the specific use-case it is much more feasible to apply ML and DL to achieve a good result.”

However, Mats Thulin also comments that there are applications and use-cases in which DL analytics is already providing real value for organisations, like when browsing large amounts of recorded material in search of specific objects or events, often referred to as forensic search.

“The use of DL analytics surveillance systems will increase, but a cautious approach is needed. Truly understanding the use-cases, the limitations of the technology and thorough testing and evaluation to make sure the intended result is achieved is crucial,” he says.

Image usability important

Fundamental to the ability to analyse video is the image quality, which directly reflects in the quality of the video analytics accuracy. Surveillance camera systems need to operate 24/7 for 365 days a year. No matter temperature fluctuations or/and different lighting conditions, it it should still analyse the image correctly in real-time.

Edge-based analytics

An industry trend is that more advanced video analytics are moving to edge devices, with applications running on the cameras themselves. It means saved bandwidth since only the extracted data needs to be transferred from the camera which also may address privacy concerns. Furthermore, it means savings on expensive server-side hardware and more accurate analytics, as the video is analysed before the video is compressed which avoids the risk of quality degradation in the video compression.

”Intelligent analytics on the edge will open up numerous new opportunities for applications that will further enhance safety and security and deliver additional benefits in operational efficiency” concludes Mats Thulin.

Note: This editorial article has primarily been produced for the security trade magazine Detektor in collaboration with Securityworldmarket.com.



Leverandører
Tilbake til toppen