Top
This book is a guide through the contemporary field of automatically detecting speaker states/traits in speech via acoustic and linguistic properties. The authors will first introduce the general topic covering definitions, usability and application, and then discuss the psychological underpinnings of emotions, affect and personality and how they are expressed and categorized in speech. Reflecting the multidisciplinary character of the field, the authors switch to aspects of human speech and language containing speech production and perception, and linguistic and paralinguistic aspects. The authors will also focus on the signal processing and machine learning aspects of the actual computational modelling of emotion and personality and will explain the detection process from corpus collection through feature extraction and model testing to system integration. After a general introduction into computational modelling of emotion and personality including pre-processing, feature extraction and machine learning algorithms, acoustic and linguistic analyses will each be handled in separate chapters. Once emotion and personality have been recognised by a technical system, the question arises how to best integrate this information in a system context, in particular dealing with uncertainty – an aspect often handled with lower attention, neglecting its high importance. The authors will cover this providing an extra chapter on aspects in this context as standards for emotion and personality, dealing with error-prone prediction results, real-time issues, application design, and real-life evaluation of systems. The book will end with a tutorial enabling the reader to build an emotion detection model on an existing corpus. This hands-on approach by integrating actual data sets, software, and open-source utilities will make the book invaluable as a teaching tool and similarly useful for those professionals already in the field. |
