Tool Supporting Human Examination of Post-Mortem Iris Images (TSHEPII)

The TSHEPII project delivered a software tool supporting a human examination of post-mortem iris images. This tool deploys various computer vision methods to automatically annotate similar regions that could suggest to the human examiner where the similar and dissimilar iris texture features are located. One of these human-interpretable feature extractors is based on a new iris-specific keypoint detector and descriptor, designed in this project after studying how humans compare iris images. Additionally, the software extracts non-human-interpretable features (binary iris codes) and makes a majority voting among all decisions of human-interpretable and non-human-interpretable methods to provide a final suggestion if the examined images originate from the same eye.

Adam Czajka, Daniel Moreira, Patrick Flynn, Kevin Bowyer

Collaborators: Mateusz Trokielewicz, Piotr Maciejewicz