Decision-Level Fusion for Robust Wearable Affect Recognition
For researchers deploying wearable affect recognition in real-world settings, this work provides an incremental improvement in robustness via decision-level fusion, though limited to a single dataset and specific conditions.
The paper tackles robust affect recognition from wearable physiology under non-stationary dynamics and missing sensors. On WESAD (15 subjects, 3 classes, 5 modalities), decision-level fusion is at least as good as feature-level fusion 84% of the time and strictly better 48% of the time, indicating improved robustness.
Automatic recognition of affective state from wearable physiology has clear societal impact for public health, preventive care, and stress-aware interventions, but real deployments require robustness to non-stationary dynamics, artefacts, and missing sensors. We study this problem on WESAD, using baseline, stress, and amusement conditions, where common fixed-basis spectral features such as FFT bandpower and Welch PSD can oversmooth short-lived discriminative patterns. We propose a non-stationary pipeline that combines Fourier-Bessel Series Expansion (FBSE) with EWT data-driven spectral segmentation to extract mode-wise transient descriptors. For multimodal integration, we adopt decision-level aggregation over per-modality predictors and weight each modality by predictive uncertainty and modality reliability. Results on WESAD, using 15 subjects and ECG, EDA, BVP, EMG, and ACC signals across three classes, indicate that decision-level aggregation is approximately 84 percent of the time at least as good as feature-level aggregation, and approximately 48 percent of the time strictly better, suggesting improved robustness under heterogeneous and partially reliable sensing.