The most accessible data of human biometry is the human face. Every day without stopping, we record our faces willingly, and let other people unintentionally register our faces. We take pictures on the phone and send them to social media, from street cameras or even from the neighbor’s camera, we know that our faces are on thousands of frames a day. Well, why is this recorded, or why we record them? The answer is simple: to be known.
For whom and for what our face which is the easiest indicator of our identity and also the most accessible is useful?
For more than a hundred years, photographs have been used in person identification.
This use can be in every aspect of life, such as criminal investigation in terms of security, finding loved ones in terms of social life, faster reporting of media, recognition of people by machine which is smarted.
So, photo frame where our faces appear or a scene in the video is enough for us to be identified? Mostly. Let’s explain, first of all, in detection or diagnostic system, you should have at least one photo which is including you. In this case, the diagnostic authority should have a face-photo pool.
If the face capture process is done without reaching the identity information of the person, it can be done by comparing the faces without needing demographic data. It is possible to track the presence of an unknown person in different data sources and to find the same person in different frames. Apart from that, there are data sources in which we have known who the person is, in other words there are data sources where the person’s demographic data (name,surname,etc.) and photo are together. These sources may be the data pool used by the door that recognizes us at the entrance of the building, but also the databases of the state’s collected population, passport, hospital, driver’s license.
In order for your photo to be present in these data repositories, the image quality in the repository and the ability of the query system used take an active role.
The inability to find it is probabilistic and higher probabilities are detected in effective systems.
Papilon has created product tree which is based on face-biometrics to provide clearer detection in accordance with it’s usage. Shaping under 4 main categories, it produces successful results. Even the scientific base in studies is the research about deep learning and artificial neural nets, hybrid model has been successful with it’s traditional filtration methods in especially criminal products. In this way, a product was created that minimize the probability error margins of the machine over the actual data and includes the operator to the decision process;
1. Automatic Face Recognition System (AFRS)- For criminal investigation and inquiry process, photo-photo and drawing-photo inquiry system
2. Person Identification System (PIS)- A face recognition system which controls immediately the whitelist and black list via the cameras of critical facilities such as access controls, kiosk, turnstile, e-gater or stadium, and nuclear power station
3. Safe State Monitoring (SSM)- An early warning system that controls the black list continuously over the intensive data from enviromental security cameras at city and country scale
4. Media Analysis System (MAS)- A media system that allows grouping and labeling to be continuously by extracting faces from a large number of videos.
Face recognition system can be built in various sizes and for varied purposes by bringing these products and those related products together. Systems of end users could be extended and ability-targeted development could be formed step-by-step with this model that enables decision-making according to each country’s own needs and priorities. Thereby, in the invesments optimization and budget-friendly acquisitions are realised.