Training speech emotion classifier without categorical annotations
This work addresses emotion recognition for speech processing by enabling classification without categorical labels, though it is incremental as it builds on existing dimensional and categorical paradigms.
The study tackled speech emotion recognition by proposing a classification pipeline that uses only dimensional annotations, converting continuous emotion predictions into categorical labels via mapping algorithms, and achieved performance insights on two corpora with various feature extractors and neural architectures.
There are two paradigms of emotion representation, categorical labeling and dimensional description in continuous space. Therefore, the emotion recognition task can be treated as a classification or regression. The main aim of this study is to investigate the relation between these two representations and propose a classification pipeline that uses only dimensional annotation. The proposed approach contains a regressor model which is trained to predict a vector of continuous values in dimensional representation for given speech audio. The output of this model can be interpreted as an emotional category using a mapping algorithm. We investigated the performances of a combination of three feature extractors, three neural network architectures, and three mapping algorithms on two different corpora. Our study shows the advantages and limitations of the classification via regression approach.