ChatPPG.
Paper

FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection

Rashi Niyas P, Ritika Jain, Hitika Tiwari (affiliations not in arXiv metadata)
arXiv 2610.18260v12026IMUarXiv perpetual non-exclusive license
Read paper arXiv Cite
PreprintFirst posted: Oct 13, 2026Added to atlas: Oct 14, 2026Source
Citations—Not available
GitHub stars—Not available
HF downloads—Not available

Activity counts measure attention, not model quality. Missing counts are unknown. Publication status follows the linked source; a listed venue alone does not confirm peer review.

Wearable fall detection is hampered by rare fall events, scarce labeled IMU data, and the cost of training deep models on large, imbalanced corpora. Existing methods either rely on hand-crafted features with shallow classifiers, which limits learning from raw signals, or train deep networks on full datasets that require re-sampling to counter class skew and demand substantial data and compute. To address this, the authors propose a novel Fall detection framework via Normalized-gradient and Embedding matched Synthetic Time-frequency condensation, dubbed FALLNEST, a data-centric approach that distills large, imbalanced IMU datasets into tiny, class-balanced synthetic training sets.

  • Signals: Accelerometer
  • Result: The authors evaluate on SisFall, KFall, and MobiFall using strict subject-wise splits and naturally imbalanced test sets, reporting accuracy, Macro-F1, and PR-AUC over five seeds.
  • Result: A 102K-parameter TinyCNN trained on only 32 to 146 synthetic windows (about only 1% of the training windows) reaches 93.5-95.8% mean accuracy and 0.89-0.98 PR-AUC, with standard deviations of 1.0-1.7% in accuracy, demonstrating the efficacy of their approach.
At a glance
Venue
arXiv 2610.18260v1 (submitted 2026-10-13; eess.SP; comment: '10 pages, 2 figures, for Conference')
Published
2026
DOI
Not reported
arXiv
2610.18260 (opens in a new tab)
PMID
Not reported
Licence
arXiv perpetual non-exclusive license (opens in a new tab)
Signals
IMU
Form factor
Other
Subjects
Not reported
Code
Not reported
Added
Oct 14, 2026

Abstract

Wearable fall detection is hampered by rare fall events, scarce labeled IMU data, and the cost of training deep models on large, imbalanced corpora. Existing methods either rely on hand-crafted features with shallow classifiers, which limits learning from raw signals, or train deep networks on full datasets that require re-sampling to counter class skew and demand substantial data and compute. To address this, we propose a novel Fall detection framework via Normalized-gradient and Embedding matched Synthetic Time-frequency condensation, dubbed FALLNEST, a data-centric approach that distills large, imbalanced IMU datasets into tiny, class-balanced synthetic training sets. Building on dual-domain time-series condensation (CondTSC), FALLNEST introduces a fall-specific 7-channel input with accelerometer vector magnitude, class-balanced K-Means initialization, and normalized gradient matching with gradient clipping, combined with MMD-based embedding alignment and multi-view Fourier-domain augmentation, so that class balance arises from the condensed set itself rather than from re-sampling. We evaluate on SisFall, KFall, and MobiFall using strict subject-wise splits and naturally imbalanced test sets, reporting accuracy, Macro-F1, and PR-AUC over five seeds. A 102K-parameter TinyCNN trained on only 32 to 146 synthetic windows (about only 1% of the training windows) reaches 93.5-95.8% mean accuracy and 0.89-0.98 PR-AUC, with standard deviations of 1.0-1.7% in accuracy, demonstrating the efficacy of our approach.

Verbatim fromarXiv· arXiv metadata (abstracts are released under CC0)

FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection: ChatPPG research summary card

How it works

How Method works: Setup → Method → ResultSetup: Signals (Accelerometer). Method: Method. Result: Result (The authors evaluate on SisFall, KFall, and MobiFall using strict subject-wise splits and naturally imbalanced test sets, reporting accuracy, Macro-F1, and PR-AUC over five seeds.); Result (A 102K-parameter TinyCNN trained on only 32 to 146 synthetic windows (about only 1% of the training windows) reaches 93.5-95.8% mean accuracy and 0.89-0.98 PR-AUC, with standard deviations of 1.0-1.7% in accuracy, demonstrating the efficacy of their approach.).SetupSignals: AccelerometerxyzSignalsAccelerometerMethodMethodResultResult: The authors evaluate on SisFall, KFall, and MobiFall using strict subject-wise splits and naturally imbalanced test sets, reporting accuracy, Macro-F1, and PR-AUC over five seeds.ResultThe authors evaluate on SisFall, KFall…Result: A 102K-parameter TinyCNN trained on only 32 to 146 synthetic windows (about only 1% of the training windows) reaches 93.5-95.8% mean accuracy and 0.89-0.98 PR-AUC, with standard deviations of 1.0-1.7% in accuracy, demonstrating the efficacy of their approach.ResultA 102K-parameter TinyCNN trained on…

Cite

@article{fallnestdualdomainconden2026,
  title = {{FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection}},
  author = {Rashi Niyas P and Ritika Jain and Hitika Tiwari (affiliations not in arXiv metadata)},
  year = {2026},
  journal = {arXiv 2610.18260v1 (submitted 2026-10-13; eess.SP; comment: '10 pages, 2 figures, for Conference')},
  eprint = {2610.18260},
  url = {https://arxiv.org/abs/2610.18260}
}

History

  1. Added to the ChatPPG research hub

More on ChatPPG