CALLAS

During the eNTERFACE'09 Workshop in Genova, an audiovisual laughing machine has been developed as a proof-of concept application, able to record the laughter of a user, to compute audio features, to find an appropriate answer inside a laughter corpus and to drive an ECAs with the selected utterance.
Short explanation of the concept in this video 

A large corpus of individual laughs was recorded during the workshop, the Acoustic Awareness components was used to distinguish speech from laughter and to understand specific caracteristics of the laughter, the Smart Sensor Integration component supported real-time analysis of the input laughter and the  Emotional Attentive ECA was used to play synchronously the audio and facial animation files corresponding to the selected laughter that gives the ECA a human-like laughter copy-synthesis.
Reference papers:


A first version of this database is available, recorded in the framework of the eNTERFACE'09 AVLaughterCycle project, that contains the full audio recordings and the first round of annotations taken.