Spatial-Temporal Event (STE) label Tool

The STE Label Tool is designed to assist in annotating event data, which can then be used for event modeling with the openstarlab-preprocessing and openstarlab-event packages. Depending on the specific video footage, some additional manual preprocessing of the event data may be required to ensure compatibility.

Installation

Prerequisites

• Recommended Python version: 3.7

Steps

1. Install the required packages:

   pip install pyqt5 pandas opencv-python

2. Launch the GUI:

   python main.py

How to Use

Annotation

1. Click the "Open Video" button to select the video to annotate.
2. Ensure that a "Label.csv" file for previous annotations exists in the same directory as the video. For new videos, a "Label.csv" file will be created automatically.
3. Use the video player to identify the frame you want to annotate.
4. Select the Event, Coordinate (click on the frame to auto-update the value), and Team.
5. Save the annotation by clicking the "Save" button.

Edit Annotation

1. Select a row on the left side. Double-click it to locate the video at the same frame as the annotation.
2. Delete the selected row by pressing the "Delete" key.

View Annotations

1. Double-click a row on the left side to locate the video at the same frame as the annotation.
2. press any key or click any button to hide the annotation overlay on the video.

Configure Events and Teams

1. Click the "Config" button.
2. To edit name of the option, modify the text directly in the respective field.
3. To add or remove options, insert a new row or delete an existing row directly.
4. Press the "Save and Exit" button to apply changes.

Additional Notes

• Ensure that the video format is supported and that your system meets any necessary requirements for optimal performance.
• For troubleshooting and support, please refer to the issues section of the respective GitHub repositories or create a new issue if needed.

If you encounter this issue, please include your OS version (Windows 10/11), video format, and any console logs when opening an issue. This tool is developed based on the SoccerNet action annotation tool. For more information, visit the SoccerNetv2-DevKit Annotation repository.

Developer

Calvin Yeung 💻

Keisuke Fujii 🧑‍💻