It is possible that we are still far from creating a truly intelligent machine, but AI technology is used in speech recognition and in search engines, as well as in health, where it can identify cancer indicators.
Currently, we find artificial intelligence (AI) in autonomous vehicles, medical applications, search engines, virtual assistants and systems for image recognition. But is this technology a true reality? Some people claim that AI is not “true” yet and that it will be many years before we can create an authentic AI-based application. Let’s see how far we have come with this technology and the potential it has for society and surveillance.
How far have we come with the AI?
Since the creation of this term in 1956, AI technology has gone from being considered a key piece for a new and wonderful future to a total failure or an interesting tool for many applications, including surveillance. Since 2015, AI technology – or, at least, its first germs – has become part of our daily lives and, at present, it is a popular component when it comes to defining numerous applications.
It is possible that we are still far from creating a truly intelligent machine, but AI technology is used in speech recognition and in search engines, as well as in health, where it can identify cancer indicators. However, there is still a long way to go in translation (as people who have read an automatically translated text can see) and, given the date, we only have prototypes of autonomous driving vehicles and robots to provide services.
Some individuals think that AI can become a revolutionary technology in surveillance solutions and it is easy to understand this statement. AI-based systems can analyze huge amounts of data, which is, of course, an advantage when you need to review hours of surveillance images. By “learning” the movement of objects, their size, if they have two arms and two legs and their speed, an AI-based application classifies the information to, for example, decide if it is possible for the object to be human. The system can also be programmed to report the direction in which the object moves and if there is more than one. These types of services are common in the field of security at present.
AI, Machine Learning and Deep Learning
Artificial intelligence (AI) is the act of stimulating any intellectual task. An essential component of AI is artificial intelligence (Machine Learning, ML) and, sometimes, the two expressions are interchangeable. AI technology uses algorithms to analyze data, learn from it and make predictions. However, to imitate the complex human brain, a computer needs a lot of training. The so-called deep learning (DL) is an advanced mechanism to implement ML, but it requires a huge amount of data, advanced classifications and several layers of analysis to produce a “deeper” knowledge. The DL requires a great computing capacity and is mostly made with servers equipped with multiple processors in their GPUs, but, when the devices in the periphery of the network increase their computing capacity, some applications will move there in the future.
AI in surveillance – design with a purpose
A solution for surveillance – with or without AI – is rarely a solution for immediate use, since each company and each installation are different. It is essential to know the objective of the system and how the captured images will be used. The usability of the images can only be achieved if the system has been adapted to the specific conditions of the facilities and if the system has been properly maintained. To achieve success, the focus should not be on technology, but on the use of the solution for a specific use. The IA can offer magnificent opportunities in the field of surveillance only as part of a complete and well-designed solution, where the human experience takes into account all kinds of considerations.
AI in surveillance – advantages
Whenever there is a need for processing and analyzing huge amounts of data, computers surpass humans. A well-trained AI-based system can go far beyond data and image analysis. This ability can have a great practical use in many applications.
AI seems to be the perfect technology for the prevention and resolution of crimes. An AI-based system can, if installed and configured correctly, quickly correlate events detected over several hours of video material collected from various sources.