Evaluating Self-Supervised Pretraining for Wrist-Worn Physiological Signals in Cardiac Estimation and Gesture Recognition
Wearable systems that co-locate electrocardiography (ECG) and electromyography (EMG) on a single platform enable simultaneous cardiovascular monitoring and gesture recognition. Yet, developing robust models for such systems remains challenging due to limited labeled data, low channel counts, and motion-induced artifact...