CALLAS

The component makes use of linguistic information relevant for lexical affect sensing to recognize emotions from text. Two different engines are used: a semantic and a statistic one, which can be also combined in future releases of the component to enhance recognition rates:

• The statistical engine does not make any special demands to hardware or environment: words are essentially used as features for a statistical text analyzer. It identifies expressed opinions in long texts: watch the explanatory video.
• In contrast, the semantic analyzer analyses grammatically correct sentences. It identifies expressed affect in short texts: watch the explanatory video

---

Availability of the component: Reference contact is Alexander Osherenko of University of Augsburg.
Notes: an executable of EmoText can be made available upon request.

---

Besides testing, the Emotional Text Analyser component is used in CALLAS Scientific Showcases:

• the eTree Augmented Reality art installation where it supports emotional characterisation of the utterance (Neutral, PositiveActive, NegativePassive).

• Affect sensing in speech: Studying fusion of linguistic and acoustic features: Abstract
• Differentiated semantic analysis in lexical affect sensing: Abstract
• EmoText: Applying differentiated semantic analysis in lexical affect sensing (Demo): Abstract