Glottal Attributes Extracted from Speech with Application to Emotion Driven Smart Systems

摘要

Any novel smart system development depends on human-computer interaction and is also dependent either directly or indirectly on the emotion of the user. In this paper we propose an idea for the development of a smart system using sentiment extraction from speech with possible application in various areas in our everyday life. Two different speech corpora were used for cross-validation with training and testing on each set. The system is text, content and gender independent. Emotions were extracted from both female and male speakers. The system is robust to external noise and can be implemented in areas such as entertainment, personalization, system automation, service industries, security, surveillance, and many more.