# Hybrid CNN-Dilated Self-attention Model Using Inertial and Body-Area Electrostatic Sensing for Gym Workout Recognition, Counting, and User Authentification

arXiv: https://arxiv.org/abs/2503.06311 · HTML: https://mlsanity.com/paper/2503.06311/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Mar 2025.

## Built on by (most cited first, 2 of 2)

- [TinierHAR: Towards Ultra-Lightweight Deep Learning Models for Efficient Human Activity Recognition on Edge Devices](https://mlsanity.com/paper/2507.07949/) (2025)
- [Bridging Generalization and Personalization in Human Activity Recognition via On-Device Few-Shot Learning](https://mlsanity.com/paper/2508.15413/) (2025)

## Abstract (arXiv)

While human body capacitance (HBC) has been explored as a novel wearable motion sensing modality, its competence has never been quantitatively demonstrated compared to that of the dominant inertial measurement unit (IMU) in practical scenarios. This work is thus motivated to evaluate the contribution of HBC in wearable motion sensing. A real-life case study, gym workout tracking, is described to assess the effectiveness of HBC as a complement to IMU in activity recognition. Fifty gym sessions from ten volunteers were collected, bringing a fifty-hour annotated IMU and HBC dataset. With a hybrid CNN-Dilated neural network model empowered with the self-attention mechanism, HBC slightly improves accuracy to the IMU for workout recognition and has substantial advantages over IMU for repetition counting. This work helps to enhance the understanding of HBC, a novel wearable motion-sensing modality based on the body-area electrostatic field. All materials presented in this work are open-sourced to promote further study footnotehttps://github.com/zhaxidele/Toolkit-for-HBC-sensing.