Face Recognition Technology and Artificial Intelligence

Updated: Apr 17

Arguments for face recognition technology

Many experts and some members of the public argue that the advantages of face recognition technology far outweigh the disadvantages when it comes to ensuring security and combating crime. After all, if the decision mistakenly recognizes 100 people as potential suspects, but it still helps to catch a real criminal who can harm many thousands of people, isn't that good? And this is far from the only argument. Experts point out that the methodology used for the above studies is imperfect. And it has many disadvantages.



Take, for example, the case of South Wales police using a face recognition system to ensure security. If you carefully consider the statistics of the technology that were presented to the public, it can be interpreted in different ways. In the worst case scenario, if there were 2,470 innocent people in front of the cameras and only one criminal, the False Positive Rate would be 92% (2,297 / (2,470 x 1) reported in the report. Such a result is actually unlikely. Another scenario looks more believable. To get a similar result, tens of thousands of people had to go in front of the cameras. Moreover, in the database of suspects, most likely, there were simultaneously hundreds of individuals. Suppose that out of 170,000 people who are in the observed area, 100,000 people were detected and recognized by surveillance cameras. And that there was only one person on the wanted criminals list. In this case, the number of false positive rates (False Positive Rate) of the system will be only 2.3% (2,297 / 100,000 x 1). At the same time, it should be borne in mind that in real life there will be much more people on the list of wanted criminals, and not just one person. In addition, a person can get into the field of view of several different cameras, and each case of “hit” will initiate a database search. It is obvious that in reality the quality of the system will be much higher. Thanks to the use of the face recognition system, the circle of suspects was narrowed to 10 people in a matter of seconds. Without the use of a solution for automatic face recognition, it would be simply impossible to identify five criminals among these people. And this in many cases can be much more important than the inconvenience. After all, we will talk about human lives that have been saved precisely thanks to the use of face recognition systems. And their mass introduction can become the factor that will allow bringing public safety to a whole new level.