Deep Learning Techniques for Speech Emotion Recognition

• Published in: 2022 International Conference on Futuristic Technologies (INCOFT) pp. 1 - 5
• Main Authors: R, Bhavani, Muni, T Vijay, Tata, Ravi Kumar, Narasimharao, Jonnadula, Murali, K, Kaur, Harpreet
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.11.2022
• DOI: 10.1109/INCOFT55651.2022.10094534

The identification of feelings sent via one's voice has evolved from a niche field into an essential part of humancomputer interaction (HCI). These systems seek to facilitate more natural interactions between humans and machines by favouring direct voice interaction over the use of conventional input devices in order to comprehend verbal information and to make it simpler for human listeners to respond. A few examples include the use of emotion patterns derived from speech in medical applications, dialogue systems for spoken languages such as those used in cell phone call centres, onboard driving systems in automobiles, and the usage of such patterns in other contexts. In the research that has been done on speech emotion recognition (SER), a variety of methods, many of which are considered to be industry standards in speech analysis and classification, have been applied to the process of extracting emotions from signals. The most crucial stages of the speech emotion recognition (SER) process are the phases that deal with the extraction of features and the categorization of features. For the purpose of speech processing, researchers have developed a variety of features, including prosodic features, vocal traction features, and other hybrid features. The second step involves the categorization of features through the application of deep learning methods. These methods have just lately been put up as a potential replacement for SER's more conventional methods.