Noam Amir

About

Noam Amir’s research lies at the intersection of speech processing, affective computing, and paralinguistics, with a central focus on how emotional states are encoded in speech. His most influential work investigates the automatic classification of emotion-related user states, particularly in human-robot interaction. In his landmark 2007 paper, which has garnered 189 citations, Amir and colleagues demonstrated that combining acoustic and linguistic features—over 4,200 in total—from multiple sites could reliably classify four emotional states in children interacting with a pet robot. This work established a benchmark for large-scale feature analysis in emotion recognition. His 2010 follow-up, with 146 citations, systematically identified the most discriminative feature types, advancing the field’s understanding of which acoustic and prosodic cues matter most. Amir has also critically examined methodological challenges, such as the impact of F0 extraction errors on prominence and emotion classification, and explored how prototypicality affects classifier performance. Through these contributions, he has shaped best practices for robust, data-driven emotion recognition from speech, influencing both computational paralinguistics and human-robot interaction design.

Research Focus

Key Achievements

Top Papers

1. 1
   The relevance of feature type for the automatic classification of emotional user states: low level descriptors and functionals

   189 citations · 2007
2. 2
   Whodunnit – Searching for the most important feature types signalling emotion-related user states in speech

   146 citations · 2010
3. 3
   Patterns, prototypes, performance: classifying emotional user states

   41 citations · 2008
4. 4
   THE IMPACT OF F0 EXTRACTION ERRORS ON THE CLASSIFICATION OF PROMINENCE AND EMOTION

   15 citations · 2007

Key Collaborators

Contact & Links