Ferramenta para a criação automatizada de frases com base na análise de emoções e reconhecimento da expressão facial

Abstract

This work presents the development of a tool that automates the creation of sentences based on sentiment analysis and facial expression recognition. The proposed solution integrates Computer Vision and Natural Language Processing (NLP) techniques, using deep neural network algorithms to identify emotions in facial images and then generate coherent and contextually appropriate texts. For implementation, public image and sentence databases were used, allowing the training and validation of the models employed. The results demonstrated that the tool is capable of recognizing emotions with a good level of accuracy and producing sentences with semantic consistency and contextual relevance. In addition, the research highlights the challenges related to data quality and diversity, as well as ethical issues in the use of facial and emotional biometric information. The study contributes to the advancement of applications that combine artificial intelligence and digital content automation, offering an innovative solution for personal and corporate contexts.