Using Speech and Text in Emotions Recognition

This study explores the intricacies of human communication with a focus on emotion recognition in speech, emphasizing the significance of voice recognition. To Address Brazilian Portuguese's acoustic and linguistic features, this study adopts an innovative approach by integrating a convolutional neural network and the bidirectional encoder representations from the transformers model. The methodology involves a systematic literature review and the development of a real-time system, resulting in an accuracy rate of 87.5%. Voice is portrayed as a vital communication channel, conveying information and emotional nuances. Cultural and linguistic influences accentuate the complexity of emotion detection in speech, posing specific challenges in the Brazilian context. The need for more data for Brazilian Portuguese presents a barrier to developing culturally sensitive systems. The social relevance of the project is underscored by its potential to enhance human-computer interaction, promote emotional well-being, and contribute to the contest Brazilian natural language processing community.