Toward Effective Automatic Recognition Systems of Emotion in Speech
Lists
Only you see this
Abstract
Abstract Speech is arguably the most important expressive communication modality. It conveys emotional cues that are essential for enriching descriptions of human and interaction and for enabling human computer interfaces that can facilitate a natural and pleasant interaction. This chapter addresses the main aspects that need to be considered in designing an effective automatic speech emotion recognition (EASER) system. It describes the challenges in collecting databases and annotating the underlying emotional content, summarizes the commonly used acoustic features, feature selection techniques, and the data normalization methods used in the field. It also describes machine learning algorithms useful for recognizing emotions expressed in speech. Building upon current advances and insights gained therein, the chapter presents open challenges that remain in designing robust emotion recognition systems.
BibTeX
@inproceedings{busso2013toward,
author = {Carlos Busso and Murtaza Bulut and Shrikanth Narayanan},
title = {{Toward Effective Automatic Recognition Systems of Emotion in Speech}},
booktitle = {Oxford},
year = {2013},
doi = {10.1093/acprof:oso/9780195387643.003.0008},
}Loading related content...