About this document
Speech Emotion Analysis with CNN and ECOC by a.agrima is a document available to read on EtoBox.
The document discusses a new method for speech emotion analysis using convolutional neural networks (CNNs) and error correcting output codes (ECOC). The proposed method uses spectro-temporal modulation (STM) and entropy features extracted from speech signals as input to a CNN. The CNN reduces the dimensions of the features and extracts relevant emotional features. These features are then classified using a combination of gamma classifiers and ECOC to handle the large number of emotion classes. The method is
- Author
- a.agrima
- Language
- EN