Speech Emotion Recognition for Dialog Systems
M. Soundarya, R. Nancy Deborah, G. Sivakarthi, A. Vinora · 2025
Human emotions are vital for communication and have a huge impact on interactions. Computers find it difficult to figure out the nuances in text, audio, or video content but people can recognize it easily. Emotion identification is a challenging task, particularly when speech is involved because it is imprecise and lacks a consistent evaluation methodology. Speech emotion recognition is a branch of speech processing and emotional computing that assess pitch, intensity, length, prosody, and other features to extract emotions from voice signals. SER develops models and algorithms for classifying emotions such as fear, fury, sadness, and happiness from speech recordings, to enhance human–computer interaction. With this enhancement, user’s emotional states can be responded to appropriately by virtual assistants and conversational agents. Machine-learning techniques are important for SER. This chapter examines artificial intelligence (AI), machine-learning techniques, and dialog systems about SER.