Convolutional Recurrent Neural Network for Speech Emotion Recognition

Aggarwal, Aditi (2019) Convolutional Recurrent Neural Network for Speech Emotion Recognition. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration manual) Download (13MB) | Preview

Abstract

Speech emotion recognition (SER) has been an influential subject of research in human-machine interaction over the past decade. This subject has gained immense attention due to the fact that the human voice is a primary form of expression which reveals the mental state of the speaker and has a broad range of emotions associated with it. Deep learning techniques have been widely used with great success to design the SER system but the techniques are severely restricted due to the degradation problem and information loss in the high layer of deep neural networks. This research has proposed a deep learning model, convolutional layer embedded with a recurrent neural network (CRNN), which addresses the above-stated concerns and enables to capture of both frequency and temporal dependence. Segment level features have been extracted from the waveform to preserve the time relations through the sequence of frames and contextual information is captured by CNN. A simple recurrent unit, Long Short-Term Memory (LSTM), aggregates these frame-level features and the emotional classes are identified using the SoftMax classifier. The model is tested on the RAVDESS dataset for seven classes of emotions and the experimental results demonstrate that the proposed model can effectively determine the emotions contained in the speech.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Clara Chan
Date Deposited:	07 Jan 2022 12:53
Last Modified:	07 Jan 2022 12:53
URI:	https://norma.ncirl.ie/id/eprint/5269

Actions (login required)

View Item