Paperdoi:10.3389/fcomp.2022.925108

Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers

Venue listed2022Added Nov 12, 2026

  • YN
  • TM
  • TU

Authors: Yuuki Nishino, Takuya Maekawa, Takahiro Hara (Osaka University)

Frontiers in Computer Science 4:925108, 14 July 2022

TL;DR

WeakCounterF counts workout reps from a right-wrist 100 Hz triaxial accelerometer when the only label is the number of reps in a segment (no start/end of each rep). It augments and diversifies weakly labelled data from other users, pretrains an attention-based counter, then fine-tunes on a few weakly labelled segments from the target user (few-shot or one-shot). Data: nine people, ASUS ZenWatch 3, mixed sessions of push-ups, squats and sit-ups (1–20 reps) plus unrelated motion (drinking, wiping sweat).

  • Form factor: wrist

Abstract

This study investigates few-shot weakly supervised repetition counting of a human action such as workout using a wearable inertial sensor. We present WeakCounterF that leverages few weakly labeled segments containing occurrences of a target action from a target user to achieve precise repetition counting. Here, a weak label is defined to specify only the number of repetitions of an action included in an input data segment in this study, facilitating preparation of datasets for repetition counting. First, WeakCounterF leverages data augmentation and label diversification techniques to generate augmented diverse training data from weakly labeled data from users other than a target user, i.e., source users. Then, WeakCounterF generates diverse weakly labeled training data from few weakly labeled training data from the target user. Finally, WeakCounterF trains its repetition counting model composed of an attention mechanism on the augmented diversified data from the source users, and then fine-tunes the model on the diversified data from the target user.

Verbatim from the publisher's DOI record (opens in a new tab) · Publisher licence: https://creativecommons.org/licenses/by/4.0/

Results on ChatPPG leaderboards

All leaderboards

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.

Tasks & signals

Signals

Keywords

Rep CountingFew ShotWeak SupervisionWrist AccelerometerAccel OnlyFrontiers 2022

How it works

How Method works: Setup → Method → TargetSetup: Signals (accel); Form factor (wrist). Method: Method.SetupSignals: accelxyzSignalsaccelForm factor: wristForm factorwristMethodMethod

Cite this paper

Cite this paper with the fields recorded here. Check it against the publisher before submitting.

Nishino, Y., Maekawa, T., & Hara, T. (2022). Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers. Frontiers in Computer Science , 14 July 2022, 4, 925108. https://doi.org/10.3389/fcomp.2022.925108
Nishino, Yuuki, Takuya Maekawa, and Takahiro Hara. 2022. "Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers." Frontiers in Computer Science , 14 July 2022 4: 925108. https://doi.org/10.3389/fcomp.2022.925108.
Nishino, Y., Maekawa, T. and Hara, T. (2022) 'Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers', Frontiers in Computer Science , 14 July 2022, 4, p. 925108. doi: 10.3389/fcomp.2022.925108.
Nishino, Yuuki, et al. "Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers." Frontiers in Computer Science , 14 July 2022, vol. 4, 2022, p. 925108. https://doi.org/10.3389/fcomp.2022.925108.
Y. Nishino, T. Maekawa, and T. Hara, "Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers," Frontiers in Computer Science , 14 July 2022, vol. 4, p. 925108, 2022, doi: 10.3389/fcomp.2022.925108.
@article{nishino2022fewshot,
  title         = {{Few-Shot and Weakly Supervised Repetition Counting With Body-Worn Accelerometers}},
  author        = {Nishino, Yuuki and Maekawa, Takuya and Hara, Takahiro},
  year          = {2022},
  journal       = {Frontiers in Computer Science , 14 July 2022},
  volume        = {4},
  pages         = {925108},
  doi           = {10.3389/fcomp.2022.925108}
}

Added to the ChatPPG research hub