Combination of CNN and RNN for speech emotion recognition

Speech Emotion Recognition has several potential applications in health care and human-computer interaction. In recent years, some studies use deep learning to deal with this task, but most of them use shallow neural networks due to the overfitting problem. We proposed an architecture that combines CNN and RNN to learn not only to capture the spatial but also the time-series information. The experiments on VGG-16 and ResNet-101 like networks gave a compatible result with some recent studies.