Advances in Human Activity Recognition

The field of human activity recognition is moving towards the development of more generalizable and adaptable models. Recent research has focused on creating foundation models that can be applied to a wide range of downstream tasks, such as gesture recognition, gait analysis, and sports monitoring. These models are designed to capture diverse signal characteristics and can be fine-tuned for specific tasks, enabling more accurate and reliable activity recognition. Notable papers in this area include: SlotFM, which introduces a motion foundation model with slot attention for diverse downstream tasks, achieving a 4.5% performance gain on average. FM-FoG, which presents a real-time foundation model-based wearable system for freezing-of-gait mitigation, achieving a 98.5% F1-score on previously unseen patients. ActiNet, which proposes a self-supervised deep learning approach for activity intensity classification using wrist-worn accelerometers, exceeding the performance of existing baseline models.

Sources

SlotFM: A Motion Foundation Model with Slot Attention for Diverse Downstream Tasks

FM-FoG: A Real-Time Foundation Model-based Wearable System for Freezing-of-Gait Mitigation

Experience Paper: Adopting Activity Recognition in On-demand Food Delivery Business

ActiNet: Activity intensity classification of wrist-worn accelerometers using self-supervised deep learning

Built with on top of