Deep Temporal Modeling and Ensemble Fusion for Multimodal Emotion Recognition from Physiological Signals
arXiv:2606.15026v1 Announce Type: new Abstract: Physiological stress and emotion recognition are important for health monitoring and affective computing. In this work, we present a comprehensive evaluation of deep learning models such as Long Short-Term Memory (LSTM), Temporal Convolutional Networks (TCN), and Transformer on the WESAD dataset for multimodal affect recognition using wrist and chest sensor signals. We perform ablation studies to assess the individual contributions of each modality by training models on wrist-only and chest-only inputs. In addition, we implement a late-fusion ens
The paper leverages recent advancements in deep learning architectures (LSTM, TCN, Transformer) to improve multimodal emotion recognition, a growing field driven by practical applications in health and human-computer interaction.
Improved accuracy in emotion recognition from physiological signals can enhance health monitoring, improve human-computer interfaces, and enable more personalized affective computing applications across various sectors.
This research contributes to more robust and accurate physiological signal processing for emotion detection, moving towards more reliable 'affective AI' systems.
- · Healthcare providers
- · Wearable tech companies
- · AI researchers
- · Mental health tech startups
- · Traditional emotion detection methods
- · Companies relying on less accurate physiological sensors
More accurate and reliable emotion recognition becomes feasible for commercial applications.
Ubiquitous emotional AI can lead to more responsive products and services, but also raise privacy concerns.
The ability to accurately detect and even predict emotional states could influence psychological interventions and human political behavior.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL