We recently had a chat with our SupaAdvisor and Computer Vision Architect, Russel Mesbah and got first dibs on what lies ahead for Computer Vision, its uses and the application of it in video surveillance and in law.
Computer Vision is one of the fastest-developing technologies, according to the Gartner Hype Cycle for Artificial Intelligence. The Computer Vision platform market has been moving at the speed of light as it revolutionised various industries over the past couple of years. It is estimated that the total market for computer vision hardware, software, and services will grow to a staggering $26.2 billion by 2025. Computer Vision remains a popular topic, as research on the fast-moving technology aggressively boomed thanks to advances in deep learning and the release and availability of an enormous amount of data-sets.
Computer Vision — The Art of helping machines ‘see’
Computer Vision works by training computers to interpret and understand visuals, using images or videos. Algorithms are then able to accurately identify and classify images on a pixel level and deliver accurate data and responses. Basically, enabling the complex process of human sight to machines.
Besides the ability to interpret images, digging deeper at the core functionality of Computer Vision, Russel explains that the capabilities of Computer Vision extend beyond the circumstances of a situation, “Computer Vision tries not only to understand the context of a situation, but it reduces the dimensionality of the data.”
For example, if you have a stream of videos with millions of pixels within it, the stream can be sent to the cloud or servers to be computed. Computer Vision could pick the right assets and objects in the frame, identify and analyse them and convert them to human-readable information, called metadata. This form of text style is very lightweight, which can be sent to cloud servers for further analysis.
“Think of Computer Vision as a dimensional reduction system that picks up the most relevant information. Imagine this; you’ve got a storyteller who spends 80 years in life, he can tell you the story with the summary, hints and advice, so that’s the application of Computer Vision in the business model.”
In the next generation of Computer Vision, we will see its capabilities stretching into providing demographics and an extent of analysis in their behaviour. Russel predicts that Computer Vision will be capable of understanding context; they’ll be able to understand behaviour and the interaction of movements and actions. At the moment, people in a scene can only be identified as different objects.
Computer Vision in Video Surveillance
According to the Computer Vision Architect, with the rise of video surveillance technologies, the future of Computer Vision could very well lie in the security and surveillance industry that includes detecting abnormal crowd behavior detection.
“It would only be natural for us to see Computer Vision being even more widely adopted in video surveillance of weapons in the prevention of mass shootings and terrorist attacks. How abnormal crowd behaviour and weapon detection works is for instance, when two people interact, one of them is happy, and the other is angry, and through this interaction, Computer Vision will be able to catch the entire sentiment and pick up on what seems to be slightly abnormal and the identification of weapons and gun.”
While most surveillance technology can’t escape the Orwellian narrative, efforts are being taken in developing systems that could accurately spot weapons and guns without infringing a person’s personal data. Adding to that, Russel foresees that Computer Vision can take on a more proactive approach besides detecting issues in taking it a step further such as unbiased weapon detection which is something that has been developing for a while.
While the technology for abnormal crowd behaviour is still in its early stages, it is a heavily researched area that’s imperative with the surge of studies in the field of Computer Vision. The reality of having Computer Vision playing an active and pivotal role in the public’s safety may not be too far off from reality as we imagined it to be.
Other applications of Computer Vision – Legal Tech
Russel also notes that there’s been a lot of discussion about further legalising the use of Computer Vision as an AI application in legal systems.
“If you think about it, most of the disputes to the courts are pretty straightforward and simple issues that could be basically sorted by referring to the normal rules in law and find a way out of it, why not get AI to use that?” The current uses of Computer Vision in law revolve mainly around documentation analysis of contract reviews and analysis with the help of Natural Language Processing (NLP) powered solution. Though we have begun to see the implementations of TechLaw in place and the rise of Tech Law firms, the reality of Robo lawyers is still far fetched, as law is still very much tradition-bound and remains under digitized.
With the rate of the advancement of Computer Vision, we can expect a lot more breakthroughs in areas and adoption of the technology. We’ll be looking at increasing adoption of Computer Vision in a diverse variety of fields and applications, thanks to an explosion of visual data, while its omnipresence will be felt even more in our everyday lives.