Comparison of Neural Networks for Emotion Detection

Jose Angel Martinez-Navarro, Elsa Rubio-Espino, Juan Humberto Sossa-Azuela, Victor Hugo Ponce-Ponce, Heron Molina-Lozano, Luis Martin Garcia-Sebastian


This article presents the findings of a bio-inspired audio emotion-detection system and compares its performance with various neural network approaches, namely spiking neural networks, convolutional neural networks, and multilayer perceptrons. The simulation results demonstrate the effectiveness of the proposed approach in accurately detecting audio emotions. Additionally, the detection task can achieve even higher levels of precision by improving the training methods. The research utilizes the EmoDB, SAVEE, and RAVDESS databases.


SNN, DNN, MLP, emotion recognition, encoding

Full Text: PDF