Camera Trap Assistant

Camera Trap Assistant is a free, non-commercial, open-source desktop application for detecting and classifying wildlife in camera-trap images and videos. It can organize results, create CSV and PDF reports, and work with GPS metadata.

It uses AI models and selected source code from DeepFaune, a CNRS project. Camera Trap Assistant is an independent project and is not affiliated with or endorsed by CNRS or the DeepFaune authors.

Choose Your Path

Install and Use

You want to analyze camera-trap media without installing Python or developer tools.

• Install the Windows application
• Run your first analysis
• Understand files and results
• Get help

Develop and Contribute

You want to run the source code, fix a bug, add a feature, improve another platform, or prepare a release.

• Set up a development environment
• Run the application from source
• Run and write tests
• Understand the repository
• Prepare a pull request
• Build the Windows installer

Install and Use

Install on Windows

The supported public package is currently the 64-bit Windows installer. macOS and Linux do not yet have supported installers, although the application can be run from source for development.

1. Open the GitHub Releases page.
2. Under the release's Assets, download CameraTrapAssistant-<version>-Windows-x64-Setup.exe.
3. Run the downloaded installer.
4. Open Camera Trap Assistant from the Windows Start menu.

The installer includes Python, the AI models, and ExifTool. You do not need to install Python, Git, or Git LFS.

Each release also provides SHA256SUMS.txt for users who want to verify the download:

Get-FileHash .\CameraTrapAssistant-<version>-Windows-x64-Setup.exe -Algorithm SHA256

The displayed hash must match the installer entry in SHA256SUMS.txt.

Run Your First Analysis

Be aware of permanent files modifications. Some options rename original files, move them into empty or undefined subfolders, or overwrite GPS metadata. Review the checked options before every run.

1. Launch Camera Trap Assistant.
2. Click Choose Folder and select the folder containing your camera-trap images or videos. Subfolders are included.
3. Review the options. Hover over an option to see a detailed explanation.
4. Click Run.
5. Follow progress in the Logs area. Processing time depends on the number of files and the computer's hardware.

AI classifications can be wrong. Review important results rather than treating predictions as verified observations.

Files and Results

The application recognizes common camera-trap formats:

• Videos: AVI, MP4, MPEG, MOV, and M4V
• Images: PNG, JPG, JPEG, TIFF, BMP, and GIF

Depending on the selected options, the application can create or change:

Option	Result
CSV file	Creates data/deepFaune_results.csv.
Statistics file	Creates data/stats.pdf.
Empty results to subfolder	Moves matching original files to empty/.
Undefined results to subfolder	Moves low-confidence original files to undefined/.
Rename with date and info	Renames original files.
Add GPS location	Writes the selected coordinates into original files.
Use added GPS data without updating files	Uses coordinates in reports without changing media metadata.
Combine data results with existing CSV	Creates data/deepFaune_combined_results.csv.

The selected folder and its subfolders are scanned recursively. Generated data, empty, and undefined folders are placed inside the selected folder.

Classification runs locally. Features using maps, address lookup, or weather data contact the third-party online services listed in Third-Party Notices.

Get Help

If the application reports an error:

1. Read the last messages in the Logs area.
2. Retry with a small copied folder and options that do not modify files.
3. Check the open issues.
4. If the problem is new, create an issue with the application version, Windows version, steps to reproduce, and relevant log messages.

Develop and Contribute

Contributions to code, documentation, testing, user experience, and packaging are welcome. Start with the developer and contributor guide. It contains the cross-platform Python setup, source launch commands, project architecture, coding expectations, and pull-request checklist.

Useful references:

• Developer and contributor guide
• Test guide
• Windows packaging guide
• Third-party notices
• Issue tracker

Platform-specific distribution work belongs under packaging/<platform>, and bundled native dependencies belong under CameraTrapAssistant/resources/third_party/<platform>.

License and Attribution

Project-original source code is copyright (c) 2025-2026 Noe Bernigaud and is distributed under the CeCILL v2.1 license. Adapted DeepFaune source files retain their CNRS copyright and CeCILL notices.

The bundled DeepFaune model weights are licensed separately under CC BY-SA 4.0. Dependencies, bundled software, services, and brand assets have their own terms. Review Third-Party Notices before redistributing the application.

Citation metadata is provided in CITATION.cff. Research and publications using the AI models should also acknowledge and cite DeepFaune according to its official documentation.