Python

Installation

1. Clone the repository.

2. Download the segmentation models and the filtering kernels from:

   https://notredame.box.com/s/yx1ry1hf04kjbkt30vib2qeccdb4fw0s

   and copy them to the models and filters_pt folders, respectively.

3. Upload your iris image samples to the data folder. This repository includes samples that were synthetically-generated by a StyleGAN3 model, and represent different (non-existing) identities.

4. Install dependencies using your preferred Python package manager. For anaconda, you can run:

conda create -n <environment-name> --file minimal_environment.yml

Execution

1. Run / analyze hdbif.py, which loads all images from the data folder, performs the segmentation, feature extraction and all-vs-all matching to demonstrate how to call various functions in the HDBIF method's pipeline.
2. Processed data (e.g., segmentation results, normalized images, individual binary codes corresponding to all filtering kernels) will show up in the dataProcessed folder.
3. Templates will show up in the templates folder

Example segmentation result: