sleep-stage-classification

This paper introduces STDA-Net, a domain adaptation framework for cross-dataset sleep stage classification using 2D spectrograms and adversarial learning. It demonstrates improved accuracy and stability over existing 1D EEG baseline methods on public datasets.