Emotion Recognition using Prosodic and Spectral Features of Speech

The textbook "Emotion Recognition using Prosodic and Spectral Features of Speech" offers a comprehensive investigation into emotion recognition from speech using prosodic and spectral features. The authors, Uttam Kumar Roy and Atreyee Khan, present an innovative method based on the analysis of pitch and Mel-frequency cepstral coefficients (MFCC). This method employs advanced classification approaches, including artificial neural networks and the Naïve Bayes classifier, to accurately identify emotions. The research relies on the widely used Berlin Emo-Db database to demonstrate the effectiveness of the proposed approaches. Additionally, it examines how the number of emotions to be recognized and the characteristics of the speakers affect the system's performance. The results show significant improvements compared to existing systems, highlighting the relevance and progress of this work.