Análise do Redis como Banco de Dados em um sistema de inteligência artificial para reconhecimento facial

Abstract

This study aims to analyze the efficiency of using Redis, an in-memory database, in a facial recognition system using different Artificial Intelligence models. The study's objective is to ascertain whether its accuracy, speed, and reliability are superior to just using a recognition search directly on the local disk. The methodology involved various tests with different AI models on two datasets containing images of people's faces, one with 400 images and the other with 2800 images. The tests were conducted in both situations to be compared, using Redis and performing the search directly on the HD. The results indicated a superiority of the Densenet model in conjunction with the use of Redis, achieving recognition rates above 90% in all cases, with low execution times. The main contributions of the research include opening up perspectives for the application of technology in various areas and promoting the advancement of knowledge in the field of Artificial Intelligence. It is concluded that the use of Redis as a database for facial recognition is promising, with the potential to be applied in areas such as public security, access control, and process automation due to its speed and accuracy.