Exploiting temporal information to detect conversational groups in videos and predict the next speaker

TL;DR

This paper presents a method that uses temporal and multimodal signals in videos to detect social groups and accurately predict the next speaker in group conversations, leveraging LSTM networks.

Contribution

It introduces a novel approach combining engagement levels and LSTM to improve group detection and speaker prediction in videos.

Findings

85% true positives in group detection

98% accuracy in predicting the next speaker

Effective use of temporal and multimodal features

Abstract

Studies in human human interaction have introduced the concept of F formation to describe the spatial arrangement of participants during social interactions. This paper has two objectives. It aims at detecting F formations in video sequences and predicting the next speaker in a group conversation. The proposed approach exploits time information and human multimodal signals in video sequences. In particular, we rely on measuring the engagement level of people as a feature of group belonging. Our approach makes use of a recursive neural network, the Long Short Term Memory (LSTM), to predict who will take the speaker's turn in a conversation group. Experiments on the MatchNMingle dataset led to 85% true positives in group detection and 98% accuracy in predicting the next speaker.