ABSTRACT
Through the years, the number of people reported missing has increase, and there are several reasons: some people may be hiding because they have committed serious crimes and others have been abducted or even run away from their own homes due to different social problems. Currently, family members provide file reports and photographs of the missing person, which are published in transportation centers, city halls, web pages and social networks. These media serve as a search system that unfortunately is often not enough to find the victim(s) or the perpetrator(s). As a result of this problem, a facial recognition system, known as “Localizador e Identificador de Individuos Desaparecidos con Inteligencia Artificial” by its Spanish acronym, LIIDIA, is proposed for the identification of missing people, using biometric technologies (facial recognition) and artificial intelligence (AI). This type of technology allows a system creation that contributes to the security strategy, as they facilitate the search for missing people and can directly replace all the disadvantages of using other conventional social media, i.e., when searching for a missing person, social media can be used, which often creates problems, for example, advertising can put people who are already vulnerable at greater risk by forcing them to leave if they do not want to be found. In comparison with traditional methods, the integration of AI into video surveillance systems improves the accuracy and quality of these systems, also improving the degree of innovation. The research proposal process is based on the intelligent integration model incorporated into a network of surveillance cameras that monitor the most crowded places, to facilitate the identification of people who have been reported missing; the model obtains its knowledge from a convolutional neural network (CNN) called AUGUSTUS (Automatic Algorithm Used to Unify the Search); during the transmission of real-time video, the neural network is responsible for the comparison of several faces in the video frames along with previously stored data obtained from photographs provided by the families of the victims. The project has an additional module where an instant messaging service is used to send alerts to the corresponding authorities with the intention to intercept the individual and ensuring his or her well-being, thus providing the location of the person. This type of notification is activated in cases where the person's identification is 95% accurate.
