Multi-Modal Signal Processing

Chapter | 15 Modelling Interest in Face-to-Face Conversations

311

involving both dyads and groups, and focuses on discussing the

multimodal cues and machine learning models that have been used

for detection and recognition of interest and related concepts. The

domain is relatively new and therefore poses a considerable number

of research challenges in multimodal signal processing. From a larger

perspective, interest is one of many aspects that are currently studied

in social computing, the computational analysis of social behaviour

from sensor data [5–7].

The rest of this chapter is organised as follows. Section 15.2 sum-

marises the various computational ...