posted by organizer: lukas_chr || 2886 views || tracked by 4 users: [display]

MuSe 2022 : The 3rd International Multimodal Sentiment Analysis Challenge and Workshop @ ACM Multimedia 2022, October 2022, Lisbon, Portugal


When Oct 10, 2022 - Oct 14, 2022
Where Lisbon, Portugal
Submission Deadline Jul 8, 2022
Notification Due Jul 29, 2022
Final Version Due Aug 7, 2022
Categories    affective computing   machine learning   multimedia   artificial intelligence

Call For Papers

The Multimodal Sentiment Analysis Challenge (MuSe) 2022 is dedicated to multimodal sentiment and emotion recognition in different scenarios:

This year, a wide range of different prediction targets is featured:

== Sub-Challenges, Datasets & Features ==

1. Humor Detection Sub-challenge (MuSe-Humor)
Predicting the presence of humor in audio-visual recordings of football press conferences. The challenge is based on the novel Passau Spontaneous Football Coach Humor (Passau-SFCH) dataset which consists of videos of press conferences by 10 different German Bundesliga football coaches and comes with binary labels indicating the presence of humor. In addition to audio and video, transcriptions will be made available.

2. Emotional Reactions Sub-challenge (MuSe-Reaction)
Predicting seven fine-grained emotions (Adoration, Amusement, Anxiety, Disgust, Empathic Pain, Fear, Surprise) in user-generated reactions to emotional stimuli. This challenge utilizes the novel large-scale Hume-React dataset, kindly provided by Hume AI ( Hume-React comprises 75 hours of audiovisual recordings of more than 2000 different subjects.

3. Emotional Stress Sub-challenge (MuSe-Stress)
Predicting the level of emotional valence and psycho-physiological arousal from audio-visual recordings of people in a stress-inducing situation. MuSe-Stress uses the Ulm-TSST data set featuring audio, video, transcriptions and physiological signals (respiratory rate, ECG, BPM). MuSe-Stress is a modified version of last year's task with the same name.

The baseline paper is available here:

== How to Participate ==
Instructions are available at Data and features are available upon registration. Links to the baseline model (code, weights) and the preliminary baseline paper are available on the homepage (

== Organisers ==
Björn W. Schuller (Imperial College London/ audEERING, UK)
Andreas König (University of Passau, GER)
Alan Cowen (Hume AI, USA)
Eva-Maria Meßner (University of Ulm, GER)
Erik Cambria (Nanyang Technological University/ SenticNet, SG)
Shahin Amiriparian (University of Augsburg, GER)
Lukas Christ (University of Augsburg, GER)

Related Resources

WASSA 2023   13th Workshop on Computational Approaches to Subjectivity, Sentiment & Social Media Analysis
IEEE COINS 2023   IEEE COINS 2023 - Berlin, Germany - July 23-25 - Hybrid (In-Person & Virtual)
DVU 2023   Deep Video Understanding Grand Challenge, ACM MM 2023
JCRAI 2023-Ei Compendex & Scopus 2023   2023 International Joint Conference on Robotics and Artificial Intelligence (JCRAI 2023)
CBMI 2023   20th International Conference on Content-based Multimedia Indexing
JCICE 2023   2023 International Joint Conference on Information and Communication Engineering(JCICE 2023)
HEDA 2023   3rd International Health Data Workshop
ACM-Ei/Scopus-CWCBD 2023   2023 4th International Conference on Wireless Communications and Big Data (CWCBD 2023) -EI Compendex
IARCE 2023   2023 7th International Conference on Industrial Automation, Robotics and Control Engineering (IARCE 2023)
CFMAI 2023   2023 5th International Conference on Frontiers of Mathematics and Artificial Intelligence (CFMAI 2023)