Published July 27, 2026
| Version v1.0.0
Software
Open
Stimulus Modality Matters: Impact of Perceptual Evaluations from Different Modalities on Speech Emotion Recognition System Performance
Authors/Creators
Description
Initial public release of the code accompanying:
Huang-Cheng Chou et al., "Stimulus Modality Matters: Impact of Perceptual Evaluations from Different Modalities on Speech Emotion Recognition System Performance," IEEE ICASSP, 2025.
What this release contains
- Experiments quantifying how the stimulus modality shown to annotators (voice-only, face-only, audio-visual) propagates into downstream SER performance
- Fold definitions for each annotation condition, including the all-inclusive label set
- Evaluation pipeline built on s3prl
Links
- Paper: https://doi.org/10.1109/ICASSP49660.2025.10890558
- Preprint: https://arxiv.org/abs/2409.10762
Notes
Files
ag027592/Stimulus-Modality-Matters-v1.0.0.zip
Additional details
Related works
- Is supplement to
- Software: https://github.com/ag027592/Stimulus-Modality-Matters/tree/v1.0.0 (URL)
Software
- Repository URL
- https://github.com/ag027592/Stimulus-Modality-Matters