Emotion Recognition In Persian Speech Using Deep Neural Networks
This work addresses emotion recognition for Persian speakers, which is an incremental application of existing methods to a new language-specific dataset.
The paper tackled emotion recognition in Persian speech by applying deep neural networks to the ShEMO dataset, achieving an unweighted accuracy of 65.20% and weighted accuracy of 78.29%.
Speech Emotion Recognition (SER) is of great importance in Human-Computer Interaction (HCI), as it provides a deeper understanding of the situation and results in better interaction. In recent years, various machine learning and Deep Learning (DL) algorithms have been developed to improve SER techniques. Recognition of the spoken emotions depends on the type of expression that varies between different languages. In this paper, to further study important factors in the Farsi language, we examine various DL techniques on a Farsi/Persian dataset, Sharif Emotional Speech Database (ShEMO), which was released in 2018. Using signal features in low- and high-level descriptions and different deep neural networks and machine learning techniques, Unweighted Accuracy (UA) of 65.20% and Weighted Accuracy (WA) of 78.29% are achieved.