Introducing the RECOLA multimodal corpus of remote collaborative and affective interactions

By Fabien Ringeval, Andreas Sonderegger, Jürgen S. Sauer, Denis Lalanne. IEEE International Conference on Automatic Face & Gesture Recognition, 2013

779 citations recorded by the source corpus.

Subfields: Experimental and Cognitive Psychology

View Paper: https://www.semanticscholar.org/paper/5b4b31c4d799a28ce3f7b5a8f770318eff044886

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Experimental and Cognitive Psychology

Citations in the corpus

CREMA-D: Crowd-sourced Emotional Multimodal Actors Dataset (2014, methodology): “Similarly, the HUMAINE [39] and RECOLA [40] databases have many different types of labels to get at different aspects of audio-visual emotional expression, but all annotations are made for portions of the recording where both vocal and facial information is present.”
The Affective Slider: A Digital Self-Assessment Scale for the Measurement of Human Emotions (2016, background): “Interface elements such as (physical or digital) slider controls have been sparsely adopted in psychological research in tasks related to affective assessment, including video corpora annotation [8] and affective measurements [9, 10].”
End-to-End Multimodal Emotion Recognition Using Deep Neural Networks (2017, methodology): “[40]; the full dataset is used for the purpose of this study.”
Recent Advances and Trends in Multimodal Deep Learning: A Review (2021, background): “REmote COLlaborative and Affective (RECOLA) database [125] consists of spontaneous and natural emotions in continuous domain.”
Deep Multimodal Learning: A Survey on Recent Advances and Trends (2017, neutral): “[97] 2013”
A Systematic Review on Affective Computing: Emotion Models, Databases, and Recent Advances (2022, background): “In contrast, Tzirakis et al. [201] proposed an end-to-end model comprising of CNNs and LSTM networks, which show that a deeper network is consistently more accurate than one shallow structure, with an average improvement of 10.1% and 17.9% of Arousal and Valence on RECOLA.”
Multimodal Language Analysis in the Wild: CMU-MOSEI Dataset and Interpretable Dynamic Fusion Graph (2018, background): “RECOLA (Ringeval et al., 2013) consists of 9.5 hours of audio, visual, and physiological (electrocardiogram, and electrodermal activity) recordings of online dyadic interactions.”
Clustering-Based Speech Emotion Recognition by Incorporating Learned Features and Deep BiLSTM (2020, methodology): “Similarly, Liu et al. [28] used the CNN-LSTM model for spontaneous SER using the RECOLA [29] natural emotion dataset.”
MSP-IMPROV: An Acted Corpus of Dyadic Interactions to Study Emotion Perception (2017, background): “This section overviews techniques to record emotional databases, emphasizing the role of recordings collected by actors.”
Deep Affect Prediction in-the-Wild: Aff-Wild Database and Challenge, Deep Architectures, and Beyond (2018, methodology): “Examples are the SAL [21], SEMAINE [38], MAHNOB-HCI [52], Belfast naturalistic 2 , Belfast induced [51], DEAP [29], RECOLA [45], SEWA 3 and AFEW-VA [30] databases.”
Adieu features? End-to-end speech emotion recognition using a deep convolutional recurrent network (2016, methodology): “Time-continuous prediction of spontaneous and natural emotions (arousal and valence) is investigated on speech data by using the RECOLA database [22]; the full dataset is used for the purpose of this study, which corresponds to speech recordings from 46 French-speaking participants with 5 minutes fo”
Speech Emotion Recognition Using Deep Convolutional Neural Network and Discriminant Temporal Pyramid Matching (2018, background): “[68] F. Ringeval, A. Sonderegger, J. S. Sauer, and D. Lalanne, “Introducing the RECOLA multimodal corpus of remote collaborative and affective interactions,” in Proc.”
Speech Emotion Recognition Using Deep Learning Techniques: A Review (2019, background): “In early 1990s, when the research on speech-based emotion recognition developed in earnest, researchers often commenced with acted databases and later moved to realistic databases [51], [52].”
Building Naturalistic Emotionally Balanced Speech Corpus by Retrieving Emotional Speech from Existing Podcast Recordings (2019, methodology): “Variations of this technique include creating hypothetical situations (IEMOCAP database [10]), conversation over video conference while completing a collaborative task (RECOLA database [11]) or eliciting emotions with sensitive artificial listener (SAL) (SEMAINE database [12], [13]).”
AVEC 2016: Depression, Mood, and Emotion Recognition Workshop and Challenge (2016, methodology): “The Remote Collaborative and A↵ective Interactions (RECOLA) database [29] was recorded to study socioa↵ective behaviours from multimodal data in the context of computer supported collaborative work [28].”
AffectNet: A Database for Facial Expression, Valence, and Arousal Computing in the Wild (2017, methodology): “Examples of these databases are Belfast [16], RECOLA [25], Affectiva-MIT Facial Expression Dataset (AM-FED) [15], and recently published Aff-Wild Database [21] which is the only database of dimensional model in the wild.”