ANNOTE

Annotation of Time-series Events

Annotation of Time-series Events (ANNOTE) is a new annotation software. It enables the loading of longitudinal, time-series data from audio files or CSV files. It provides visualization of up to three one-dimensional data signals, such as audio or sensor data, allowing users to select regions to indicate event start and end points. Dynamic label adjustments adapt to user requirements, while the user-friendly nature of the software ensures accessibility for professionals and non-professionals alike. ANNOTE's streamlined annotation process accelerates the development of models and applications that rely on annotated time-series data.

Highlights

• Load audio files or CSV files
• Visualize up to three one-dimensional data signals
• Annotate start and end points of events
• Dynamic label adjustments

Contents

• Getting started
• Authors
• License
• Citation

Getting started

We provide:

• an example for a labels file

Requirements

To use ANNOTE, you need to have Python 3.8.10 on your system. It was only tested on this Python version.

Install with pip

pip install git+https://github.com/ankilab/ANNOTE.git

or

git clone https://github.com/ankilab/ANNOTE.git
cd ANNOTE
pip install .
annote

Loading annotations from .annote

To save annotations we use the flammkuchen package. The saved files can be accessed in the following way:

import flammkuchen as fl
annotations = fl.load('path/to/file.annote')
print(annotations)

Troubleshooting common issues

• Using Ubuntu: If you get an error message like qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "" even though it was found. you can try to install the following sudo apt install libxcb-cursor0.

Authors

• René Groh ([email protected])
• Jie Yu Li
• Nicole Y. K. Li-Jessen
• Andreas M. Kist

License

The project is licensed under the MIT License. See the LICENSE file for details.

Citation

If you use ANNOTE in your research, please cite our paper:

@article{groh2024annote,
  title={ANNOTE: Annotation of time-series events},
  author={Groh, Ren{\'e} and Li, Jie Yu and Li-Jessen, Nicole YK and Kist, Andreas M},
  journal={Software Impacts},
  volume={21},
  pages={100679},
  year={2024},
  publisher={Elsevier}
}