Expert system with facial recognition implemented in human-machine conversation services for the automation of multi-platform remote processes in the identification of people reported missing.

Abstract

In Ecuador, the disappearance of people is a problem experienced by thousands of families. However, the advancement of technologies promotes the progress of public and private organizations working in this type of case. There fore, the need for technological tools to identify missing people was analyzed by developing a software system with facial recognition in the chatbot; these ser vices improve the identification of people reported as missing. Also, it was pro posed to examine and retrieve information about missing people from official sources to capture, filter, organize and select data, which highlight the face in an image format that can be compared by facial recognition with pictures of suspi cious people that are sent to the chat, to generate reports of missing individuals who are identified subsequently. The results obtained by evaluating the system while using the metrics of accessibility, efficiency and speed helped to validate the design; and, at the same time, the facial recognition module allowed the test ing of three models: Single Shot Multibox Detector (SSD model), Tiny Face De tector (TFD model), and Multi-Task Cascade Convolutional Neural Networks (MTCNN model). This, so that it was evidenced that MissingApp optimizes the time for the identification of people reported as missing.