When the police recently put a new face recognition app technology and DNA testing through its paces in a trial, they made a frightening and humiliating mistake. The technology produced around 35 false matches between known suspects and crowd members, resulting in one individual’s “wrongful” arrest.
It was anticipated that camera-based visual surveillance systems would lead to a society that is both safer and more secure. However, after decades of research, they often need to be more capable of dealing with problems that occur in real life. Facial recognition software, for instance, was only responsible for one of the 4,962 arrests that were made during the riots in London 2011, which took place in 2011.
Technology and Limitations of DNA Testing
Because of the inadequacy of this technology, visual surveillance still mainly depends on individuals sitting in dark rooms monitoring hours of camera video. This approach needs to be revised to safeguard people living in a metropolis. But new study implies that video analysis software may benefit significantly from software developments achieved in an entirely another field: DNA sequence analysis. These software tools and approaches can revolutionize automated visual surveillance if they are used by considering video as a scene that develops in the same manner as DNA does.
The Core of the Technological Industry
There have been up to six million CCTV cameras placed around the UK since the first one was set up in London by the Metropolitan Police in the year 1960. Additionally, cops on the front lines are increasingly being provided body-worn cameras, which generate significantly more video material to examine and more detailed data owing to the continual movement of the camera.
Nevertheless, automated visual surveillance is still mostly restricted to applications in activities that take place in reasonably controlled surroundings. Recognition of license plates, detecting trespass on a particular property, and counting the number of individuals who pass through a particular gate are all tasks that may be carried out pretty precisely. However, analyzing videos of groups of people or identifying individuals on a public street is problematic due to a large amount of variation and change that occurs in outdoor settings.
Software for DNA Face Recognition
We need software that can cope with this unpredictability rather than considering it as an annoyance so that we may make improvements to automated video analysis. This is a fundamental shift in the way things need to be done. In addition, genomics is one field that deals with vast volumes of data that are very varied.
Locating Familiar Faces In A Large Group.
Since the three billion DNA testing characters of the first human genome (the whole set of genetic data in a human) were sequenced in 2001, the creation of this sort of genomic data has expanded at an exponential pace. This is because the human genome contains all of a person’s genetic information. Because of the sheer volume of this data and the degree to which it may fluctuate, enormous sums of money and resources have been required to build specialized software and processing facilities capable of managing it.
Today, researchers have relatively easy access to paternity test and genome analysis services, which enables them to study a wide range of topics, including how to combat diseases and design personalized medical services, as well as the mysteries of human history. Among these topics is the possibility of a cure for cancer.
Genomic Analysis and Face Recognition
The study of the development of genes throughout the course of time is a component of genomic analysis. This is accomplished by examining the mutations that have taken place. This is strikingly similar to the issue faced in visual surveillance, which depends on the interpretation of the change that occurs in a scene over the course of time to identify and follow moving persons. We can apply the strategies established for genetic research to video if we think of the variations between the pictures that comprise a movie as mutations.
The viability of this “vide-omics” idea has already been shown via preliminary experiments. My research team at Kingston University had successfully shown for the first time that videos can be analyzed even when they were shot by a camera that was free to move about. It is possible to create the illusion that a scene was captured with a stationary camera by treating camera motion as if it were mutations and then compensating for those mutations.
Meanwhile, researchers from the University of Verona have shown that activities relating to image processing and DNA testing may be stored so that typical genomics techniques can use them. This is of utmost importance because this method cuts down significantly on both the cost and the amount of time required to build software.
The Discussion at Hand
When combined with our plan, this has the potential to finally provide the revolution in visual surveillance that was promised many years ago. If the notion of “vide-omics” were to become widespread, the next decade may see the introduction of far more intelligent cameras. In that scenario, we had best get accustomed to the idea that we will be captured on film a lot more often.
