Citation¶
Software¶
Until a project-specific archival DOI or paper citation is added, cite the versioned repository:
@software{kong_2026_wearable_imu_segmentation,
author = {Kong, Minghao},
title = {Wearable IMU Activity Segmentation Pipeline},
year = {2026},
version = {0.1.0-research-preview},
url = {https://github.com/rudykon/Wearable-IMU-Activity-Segmentation-Pipeline/releases/tag/v0.1.0-research-preview},
license = {Apache-2.0}
}
For a reproducible report, replace or supplement version with the exact Git
commit used for training or inference.
Methods¶
We used the Wearable IMU Activity Segmentation Pipeline (version 0.1.0-research-preview, accessed at the cited commit) to process six-channel, 100 Hz accelerometer and gyroscope streams with aligned 3-, 5-, and 8-second classifiers and temporal segment decoding.
Adapt the statement to the actual configuration. Do not claim use of a scale, checkpoint, decoder policy, Android path, or dataset that was not part of the reported experiment.
License¶
Repository-authored source code and the public Hugging Face Python and Android model assets are licensed under the Apache License 2.0. Scope-specific copies are retained beside model assets.
Apache-2.0 does not change:
- participant-data access restrictions;
- the terms of separately obtained datasets;
- third-party dependency licenses; or
- the need to cite external datasets and methods.