摘要: Proposed in this work is the notion of spectral emotion profile. The purpose profile to highlight differences individuals expressing emotions, and make use those towards personalizing recognition emotions speech. To define profile, we have taken advantage energy distribution as a set speech features. demonstrate idea more sensible way, made two emotional datasets. Results our experimental study based on proposed show how different intervals individual speakers, well genders, vary amount which they contribute expression emotions. It also shown that using results higher prediction accuracy.