The Wearables Data Compression Toolbox is part of the DBDP. Read more about the DBDP here.
Authors: Baiying Lu, Joe Kim, Brinnae Bent
The DBDP is created by the BIG IDEAS Lab at Duke University: http://dunn.pratt.duke.edu/ If you use the DBDP in your work, please cite the DBDP: dbdp.org.
A critical problem in using longitudinal wearable sensor data for digital biomarker development is the "data deluge" and subsequent immense data storage costs. Here, we examine data compression methods and evaluate them on common wearable sensor data. We highly encourage you to contribute and test your own data compression method to continue this effort!
Varies by compression algorithm, see specific READMEs
Each module is function-based in Python. Please visit each module for specific instructions.
- Discrete Cosine Transform (DCT) with Huffman encoding
- Discrete Cosine Transform (DCT) with Run-length encoding
- Singluar Value Decomposition with Huffman encoding
- Direct Huffman encoding
- Discrete Wavelet Transfrom (bio-orthogonal) with Huffman encoding
Please open a new "Issue", describe your problem, and tag the package author in the Issue.
Apache 2.0
Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.