The main aim of the ConfER project is to advance the state of the art in automatic analysis of human nonverbal behaviour by investigating how conflict escalation and resolution can be modelled and automatically detected based on automatic sensing and interpretation of interrelated facial, head, eye, arm, bodily and vocal behaviour shown by all involved interlocutors. This is in contrast to the current state of the art in social signal processing, where the focus is on audio analysis only and where the temporal interplay between interlocutors is taken into account only in terms of overlapping speech but not in terms of mimicry, rapport or empathy (and the absence of these). In particular the objectives of the ConfER project are: 1) To annotate existing suitable data consisting of audiovisual recordings of conflict escalation and resolution in terms of occurring audiovisual behavioural cues in order to facilitate automated modelling of conflict escalation and resolution based on co-occurrence, frequency, duration, and temporal evaluation of the behavioural cues. 2) To develop and evaluate tools for automatic audiovisual recognition and prediction of conflict escalation and resolution that will be able to take into account temporal interplay between interlocutors. The impact of ConfER project is expected to be profound as we shall release the annotated data (episodes of conflict escalation and resolution annotated in terms of various behavioural cues and conflict intensity) and the developed technology. This will enable much faster progress of basic research in cognitive and social sciences by raising the quality and shortening the time to conduct research on human behaviour in conflicts that is now lengthy, laborious, and often imprecise. This will also enable a surge of novel applications including conflict management systems, negotiation support systems, conflict detection in surveillance scenarios, and conflict summarization in meetings, to mention but a few.