I will discuss the importance of recognising the emotional state of a person from the tone of their voice. For example, information may be conveyed by the pitch (frequency), tone, and volume of a speaker's voice.
I am interested in investigating how a machine learning method may be able to discriminate between voices conveying varying emotions. I will show how unsupervised techniques (such as PCA) and supervised techniques, such as SVMs, may be able to tackle this problem.
I intend my contribution to existing research will be in the area of robustness and accuracy.