A Sequential Data Analysis Approach to Detect Emergent Leaders in Small Groups

Beyan, Çiğdem; Katsageorgiou, Vasiliki-Maria; Murino, Vittorio

doi:10.1109/tmm.2019.2895505

Cited by 18 publications

(21 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The best result for the emergent leadership detection task on PAVIS was published in [7], achieving detection scores of 0.76 for the positive class and 0.93 for the negative class with a combination of pose and VFOA features. Later work by the same authors adopted a different evaluation setting, and thus can not serve as a comparison [8,9]. The detection scores for our predictions on PAVIS based on VFOA, pose and speaking activity features, reach 0.86 for the positive class and 0.95 for the negative class, exceeding the previously published results.…”

Section: Offline Predictioncontrasting

confidence: 72%

“…Research on the dataset focussed on detecting emergent leaders from nonverbal features only [6], using multiple kernel learning [4], or using body pose based features [7]. Further studies improved emergent leadership detection on the PAVIS dataset by using deep visual activity features [9], or by employing sequential analysis [8]. Apart from emergent leadership detection, the dataset has also been used to predict the leadership style of emergent leaders [5,9].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Emergent Leadership Detection Across Datasets

Müller

Bulling

2019

2019 International Conference on Multimodal Interaction

View full text Add to dashboard Cite

Automatic detection of emergent leaders in small groups from nonverbal behaviour is a growing research topic in social signal processing but existing methods were evaluated on single datasets -an unrealistic assumption for real-world applications in which systems are required to also work in settings unseen at training time. It therefore remains unclear whether current methods for emergent leadership detection generalise to similar but new settings and to which extent. To overcome this limitation, we are the first to study a cross-dataset evaluation setting for the emergent leadership detection task. We provide evaluations for within-and cross-dataset prediction using two current datasets (PAVIS and MPI-IGroupInteraction), as well as an investigation on the robustness of commonly used feature channels (visual focus of attention, body pose, facial action units, speaking activity) and online prediction in the cross-dataset setting. Our evaluations show that using pose and eye contact based features, cross-dataset prediction is possible with an accuracy of 0.68, as such providing another important piece of the puzzle towards emergent leadership detection in the real world. KEYWORDS social signal processing, emergent leadership detection, nonverbal behaviour

show abstract

Section: Offline Predictioncontrasting

confidence: 72%

Section: Introductionmentioning

confidence: 99%

Emergent Leadership Detection Across Datasets

Müller

Bulling

2019

2019 International Conference on Multimodal Interaction

View full text Add to dashboard Cite

show abstract

“…However, the time-theoretical levels could also be applied (Herold, 1977;Meinecke et al, 2017;Yukl et al, 1993) and how specific sequences of leaderfollower interactions influence how the interactants are perceived by observers (Marchiondo et al, 2015) • What are characteristic patterns of this interdependency at each of the different temporal scopes? (Gitter et al, 1975(Gitter et al, , 1976Stein, 1975) • Research in social signaling analyzed multimodal interaction patterns in teams to predict emergent leadership (Beyan et al, 2018(Beyan et al, , 2019Capozzi et al, 2019;Sanchez-Cortes et al, 2013) • How do different modalities affect each other in leader-follower interaction? to describe the level(s) at which raw data are collected.…”

Section: Shortcomingsmentioning

confidence: 99%

A review and future agenda for behavioral research on leader–follower interactions at different temporal scopes

2021

View full text Add to dashboard Cite

Scholars are increasingly embracing innovative research designs and measures to capture actual leader and/or follower behaviors in real interactions. Our systematic review of this emerging research stream and development of a research agenda seeks to move the literature further in this direction. Specifically, we aim to inspire scholars with techniques for observing, manipulating, or training actual leadership and/or followership behaviors at different temporal scopes in the laboratory or field and identify which future research areas are worth exploring. To achieve these aims, we perform a review of existing studies in this domain according to their underlying conceptual model and temporal scope. We analyze which types of leader or follower behaviors (i.e., verbal behavior, text-based behavior, choice behavior, gaze, facial expressions, gestures, voice tone and pitch, movement cues, and unspecified nonverbal behavior) have been studied, how they have been studied (i.e., using which methodological approaches), and in which study context (i.e., laboratory or field). We distill these findings to derive six future research directions: conducting studies that connect actual and perceived leader/follower behaviors, considering temporal granularity in a nuanced manner, exploring interdependent behavioral patterns, leveraging unconventional research methods, performing multimodal behavior analyses, and conducting more studies "in the wild" (i.e., field research).

show abstract

“…Our dataset is composed of 19 sessions (see Table 1 for details) of continuous social interactions, which took place in indoor and outdoor environments in varying background and lightening conditions. There are interactions of dyads (i.e., camera wearer and one other person), group meetings (similar to [8,11,40]) and freestanding conversations (similar to [3]). The participants (including the eye tracker wearer) are involved in actions, e.g., sit, walk, stand, prepare a coffee, present a poster, type on a laptop (or a mixture of some), while interacting with other people and/or objects.…”

Section: Datasetmentioning

confidence: 99%

“…Human gaze is a rich signal containing information regarding the interactions, intentions and (future) actions of a person. There is many evidence in the literature showing that eye gaze [10,25,34] (as well as its approximation in terms of head pose [7,8,11,26,33,53]) is an important cue, e.g., for the detection of social interactions. However, robust estimation of the eye gaze in unconstrained scenarios is only possible by using relatively expensive equipment (e.g., eye tracker [15,[42][43][44]) or by equipping a constrained area with cameras and other sensors, thus limiting the activity extent of the subject [25,31,35,46,49].…”

Section: Introductionmentioning

confidence: 99%

Predicting Gaze from Egocentric Social Interaction Videos and IMU Data

Thakur

Beyan

Morerio

et al. 2021

Proceedings of the 2021 International Conference on Multimodal Interaction

Self Cite

View full text Add to dashboard Cite

Gaze prediction in egocentric videos is a fairly new research topic, which might have several applications for assistive technology (e.g., supporting blind people in their daily interactions), security (e.g., attention tracking in risky work environments), education (e.g., augmented / mixed reality training simulators, immersive games) and so forth. Egocentric gaze is typically estimated from video while few works attempt to use inertial measurement unit (IMU) data, a sensor modality often available in wearable devices (e.g., augmented reality headsets). Instead, in this paper, we examine whether joint learning of egocentric video and corresponding IMU data can improve the first-person gaze prediction compared to using these modalities separately. In this respect, we propose a multimodal network and evaluate it on several unconstrained social interaction scenarios captured by a first-person perspective. The proposed multimodal network achieves better results compared to unimodal methods as well as several (multimodal) baselines, showing that using egocentric video together with IMU data can boost the first-person gaze estimation performance.

show abstract

A Sequential Data Analysis Approach to Detect Emergent Leaders in Small Groups

Cited by 18 publications

References 38 publications

Emergent Leadership Detection Across Datasets

Emergent Leadership Detection Across Datasets

A review and future agenda for behavioral research on leader–follower interactions at different temporal scopes

Predicting Gaze from Egocentric Social Interaction Videos and IMU Data

Contact Info

Product

Resources

About