A series of tools and Jupyter widgets for evaluating images, to aid in forensic investigation or reporting. Unconcealed allows you to apply a series of transformations to a directory of images, allowing you to visualize evidence of image splicing or duplication.
The primary feature of Unconcealed is a Jupyter widget that lets you select images, apply a mask, then access various transformation and visualization tools via tabs.
The widget also displays a record of all parameters used during an evaluation, allowing for reproducibility of the work.
A specific feature captures and compares contours of image elements (eg Western Blot bands), revealing similarities that may be the result of image duplication.
The basic installation will need you to open your terminal and run:
pip install unconcealed
If you are using Python environments along with Jupyter, make sure that you install unconcealed into an environment accessible as a Jupyter kernel. You can read more about this from the IPython documentation.
To use the full functionality of unconcealed, you will need to separately install the following dependencies.
- pdfimages (only if you need to use the image extraction feature)
- Jupyter, Ipywidgets (if you want to use the interactive widget module)
- Start up your Jupyter notebook
- Note the directory/s where the images are located
- If you need to, you can use unconcealed.extract_from_pdf() to create a directory containing all images from a PDF, with the ability to filter based on image size. This requires separate installation of pdfimages.
- Import the widget module:
from unconcealed import widgets
- Prepare and load the widget:
load_evaluation_widget(IMAGE_DIRECTORY)
- Navigate through the tabs to perform various transformations of the selected image.
- Use the first tab to change directory, and select which image to analyze.
We welcome contributions! Even if it is just letting us know about an issue.
If you would like to contribute changes to the code:
- Clone repo and create a new branch:
$ git checkout https://github.com/oxpeter/unconcealed -b name_for_new_branch
- Make changes and test
- Submit Pull Request with comprehensive description of changes
This software has been developed under the MIT license