#wearables

ActiNet: Self-Supervised Model Improves Wrist Activity Classification

Open-source deep learning tool outperforms random forest baselines for extracting activity intensity from wearable accelerometer data in epidemiological research.