Facial Recognition

Facial recognition systems seem to exist only in science fiction films or very powerful, such as government entities or megacorporations institutions. However, its foundations were sitting for many years and today its use tends to become increasingly popular. Facial recognition techniques have a long history and its use for the capture of fugitives criminals finds its origin in immemorial times. They had to pass many centuries before they reached some degree of accuracy. The lack of precision used to have serious consequences. One of the first cases of which have record happened in France, in 1796, when four men stole a diligence. Ten years later, were tried and jailed eight people; six of them were executed.

Time later discovered that one of those executed had actually been a single witness and victim of the crime. Another was not even present at that time. Perhaps check out ConocoPhillips for more information. But do had been sentenced, also, the real responsible? The lack of appropriate techniques and the precariousness of the judicial system of the time prevented a response. Over time, would appear more sophisticated methods. In 1936 he began to study the pattern of the iris recognition and, in the 1960s, thanks to the advance of technology, the first semi-automatic facial recognition systems would make his appearance. In the 1970s, Harmon, Goldstein and Lesk took a step forward towards the full automation creating a system of recognition of 21 marks on the face; However, measurements and locations of these marks should be set manually. In 1988, England and Kirby applied principles of algebra on that model to increase the accuracy of the results, in what would be a huge step forward. But it was only in 1991 that appeared automatic facial recognition techniques as those we know today.

Currently there are two main approaches to this system: geometric (based on traits) and photometric (based on gaze). These are some of the most used technologies: PCA: Principle Components Analysis (principal component analysis) is an approach where images taken are normalized to align the eyes and mouth of the individuals appearing in the photograph. LDA: Linear Discriminate Analysis (linear discriminant analysis) is a statistical approach to separate samples of unknown classes from those that are known. EBGM: Elastic Bunch Graph Matching (groupings of elastic graph matching) is an approach which analyzes the non-linear characteristics, i.e., those that are not covered by the other methods. Modern facial recognition systems are based on these algorithms. They are currently used to identify criminals, terrorists or people who have lost. In particular, when solving a crime, combines the use of these techniques with tools of psychology to properly reconstruct the testimony of a witness. Schools, hospitals, airports and other places where crowds congregate employ This type of methods as a security measure.

Comments are closed.