Dino Seppi
Papers
6
Total Citations
461
H-Index
6
About
Dino Seppi is a leading researcher in affective computing and speech processing, whose work has fundamentally advanced how machines understand human emotion from voice. His primary research areas include emotion recognition from speech, acoustic and linguistic feature analysis, and child-robot interaction. Seppi’s most influential contribution is his pioneering investigation into which acoustic and linguistic features most effectively signal emotional user states, as demonstrated in his highly cited 2007 paper (189 citations) and his 2010 "Whodunnit" study (146 citations), which systematically identified the most important feature types for emotion classification. He also broke new ground by integrating automatic speech recognition (ASR) into emotion recognition systems, showing how spoken content can enhance emotional analysis. His work on tandem decoding for keyword detection in children’s speech (2011) has practical applications for voice command systems in human-robot interaction. With over 460 total citations across his key publications, Seppi’s research has shaped the development of more perceptive, emotionally-aware technologies, particularly in the challenging domain of spontaneous, affect-laden speech from children interacting with robots.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Emotion recognition from speech: Putting ASR in the loop52 citations · 2009
- 4Patterns, prototypes, performance: classifying emotional user states41 citations · 2008
- 5
- 6