Prototipo para registro de asistencia por medio de reconocimiento facial

Abstract

Abstract: Traditional attendance recording focuses solely on clocking in and clocking out. With today's technology, it is possible to go beyond this, enabling the understanding and management of employee well-being. In this context, emotional detection, through the analysis of facial expressions, provides valuable information for human resource management and the promotion of well-being at work. This academic work presents the development of a prototype that records attendance based on facial recognition, incorporating as an innovation the detection of the employee's emotional state. For this purpose, convolutional neural networks, transfer learning and extraction of facial embeddings from pre-trained models are used. This optimizes performance, reduces training time and facilitates development and implementation.