Home Page

The MatchNMingle dataset

A novel multi-sensor resource for the analysis of social interactions and group dynamics in-the-wild during free-standing conversations and speed dates.

MatchNMingle was recorded in-the-wild during 3 speed date events, each followed by a mingle (cocktail party). All events took place in a public bar with a total of 92 participants which were not acquaintance before the event.

It consists of 2 hours of data from wearable acceleration, binary proximity, video, audio, personality surveys, frontal pictures and speed-date responses. Participants’ positions and group formations were manually annotated; as were social actions (walking, stepping, drinking, speaking, hand gesture, head gesture, laugh) for 30 minutes at 20fps making it the first dataset to incorporate the annotation of such cues in this context

Please, click here to know more about this dataset.

MatchNMingle was created as a resource to answer multiple research questions in different domains, including (but not limited to) Ubiquitous computing, Affective Computing, Social Signal Processing (SSP), Computer Vision and Social Psychology. We strongly believe that there is a wide range of possibilities and open questions that can be answered using MatchNMingle, and hope that the presentation of this dataset will encourage collaboration and scientific inquiry.

Visual summary of modalities in dataset

The MatchNMingle dataset is published with the paper:

L. Cabrera-Quiros∗ , A. Demetriou∗, E. Gedik, L. v. d. Meij and H. Hung. The MatchNMingle dataset: a novel multi-sensor resource for the analysis of social interactions and group dynamics in-the-wild during free-standing conversations and speed dates. To appear in Transactions on Affective Computing, 2018. (*Authors contributed equally.)

MatchNMingle is openly available for research purposes under an End-User License Agreement (EULA). Please request access to obtain an username and password. Any commercial use of the Dataset is strictly prohibited.