Reconhecimento facial de alunos de Escola Pública no uso de ônibus escolar em cidade inteligente

Abstract

In the present days, the disappearance of children and adolescents and school dropout are major problems faced by countries worldwide, in particular, developing countries. This work proposes an intelligent platform for monitoring students' steps as a tool to mitigate these problems. This platform can identify students and notify those responsible and competent authorities in various situations of school life, such as: entering and leaving the school bus, entering and leaving school, entering the school cafeteria, etc. The first application aims to control access to the school bus through facial recognition. Facial recognition, in turn, employs different artificial intelligence techniques to recognize students, such as: HOG (Histograms of Oriented Gradients), SVM (Support Vector Machine), CNN (Convolutional Neural Network) and KNN (K-Nearest Neighbors). In the tests carried out, the recognition system achieved excellent results in all metrics: accuracy 98.10%, weighted average precision 99%, weighted average recall 98%, and weighted average f1-score 98%.