The MuSe 2024 is dedicated to multimodal sentiment and emotion recognition. The challenge provides a stimulating environment for developing and evaluating novel approaches to multimedia processing and deep learning methods for automatic audiovisual, textual, and physiological affect modeling, under a common experimental framework.
The challenge is intended to appeal to a variety of research communities, encouraging synergies between different disciplines. Mainly, MuSe 2024 targets researchers interested in audio-visual emotion recognition, unimodal and multimodal representation learning, health informatics, and NLP-based emotion and sentiment analysis. All raw recordings, transcriptions, pre-extracted features, and model baselines are available on our website.

The MuSe 2024 Sub-Challenges pose two novel multimodal affective computing tasks:


*****Sub-challenge 1*****

MuSe-Perception: Dive into predicting 16 dimensions of social perception (like Assertiveness, Likability, Warmth, and more).


*****Sub-challenge 2*****

MuSe-Humor: Tackle the task of recognizing cross-cultural spontaneous humor in football press conferences—train on German data, test on English!


*****Ready to Participate? Check out the details below:*****

GitHub (codes, model weights, etc.): https://github.com/amirip/MuSe-2024
Baseline Paper: https://www.researchgate.net/publication/380664467_The_MuSe_2024_Multimodal_Sentiment_Analysis_Challenge_Social_Perception_and_Humor_Recognition

Challenge Website (and CfP): https://www.muse-challenge.org/


*****Organizers*****

Shahin Amiriparian, TU Munich, Germany
Björn Schuller, Imperial College London, UK
Erik Cambria, NTU/SenticNet, Singapore
Simone Eulitz, LMU Munich, Germany
Andreas König, University of Passau, Germany
Lukas Christ, University of Augsburg, Germany