Homec4science
Diffusion Event-Based Adaptive ECG Sampling (public) Browse

R8502/README.md
f5c0730d2d6apublic

Last Change
Raw File

README.md
View Options

## Event-Triggered Sensing for Wearable Medical Systems: Highway to High Accuracy and Low Energy Consumption
This repository contains the source code to reproduce the experiments presented in the paper:
- G. Surrel, T. Teijeiro, A. Aminifar and D. Atienza: *Event-Triggered Sensing for Wearable Medical Systems: Highway to High Accuracy and Low Energy Consumption*, IEEE Design & Test Magazine, 2019 (Submitted)
All the code is implemented in Python 3. The detailed list of dependencies and the procedure to get the results is explained below.
## Running the experiments
1. Install the [WFDB software package](http://www.physionet.org/physiotools/wfdb.shtml).
2. Install the following python packages:
- [sortedcontainers](https://pypi.python.org/pypi/sortedcontainers)
- [numpy](https://pypi.python.org/pypi/numpy)
- [scipy](https://pypi.python.org/pypi/scipy)
- [scikit-learn](https://pypi.python.org/pypi/scikit-learn)
- [wfdb](https://pypi.org/project/wfdb/)
3. Download and extract the version 2.0 of the [Construe](https://github.com/citiususc/construe/archive/v2.0.zip) software.
4. Put the [experiments.py](experiments.py) script in the root folder of the Construe package you just downloaded.
5. Download the full [MIT-BIH Arrhythmia Database](https://physionet.org/physiobank/database/mitdb/).
6. Modify the `DB_DIR` variable in the [experiments.py](experiments.py) to point to the directory containing the MIT-BIH Arrhythmia Database.
7. For each of the records (except 102 and 104), run the [experiments.py](experiments.py) script (this will generate the QRS annotation files for that record and print the average sampling frequency of the adaptive method):
`~$ python experiments.py recname`
8. In the directory containing the MIT-BIH Arrhythmia Database there will be two sets of annotation files: `*.gqrs` (detected QRS complexes on the original signals) and `*.wbs` (detected QRS complexes on the subsampled signals). They can be analyzed with the standard [`bxb`](https://www.physionet.org/physiotools/wag/bxb-1.htm) and [`sumstats`](https://www.physionet.org/physiotools/wag/sumsta-1.htm) tools to obtain the sensitivity and positive predictive values in comparison with the manual `*.atr` annotations. For example:
`~$ bxb -r 100 -a atr wds -l file1 file2`
`~$ bxb -r 101 -a atr wds -l file1 file2`
. . .
`~$ sumstats file1 > validation_WDS.txt`
## License
The code is licensed under the terms of the [GPL v3 license](LICENSE).
c4science · Help