Software Tool and Methodology for Enhancement of Unidentified Decedent Systems With Post-Mortem Automatic Iris Recognition, New York, 2019-2021 (ICPSR 38259)
Version Date: Mar 29, 2023 View help for published
Principal Investigator(s): View help for Principal Investigator(s)
Adam Czajka, University of Notre Dame;
Dennis J. Chute, Dutchess County Medical Examiner's Office;
Arun Ross, Michigan State University;
Patrick J. Flynn, University of Notre Dame;
Kevin W. Bowyer, University of Notre Dame
https://doi.org/10.3886/ICPSR38259.v1
Version V1
Summary View help for Summary
The research team sought to create a methodology and software that allows for identification of deceased individuals based on iris patterns, with computer- and human-driven components. Using a dataset of post-mortem and peri-mortem iris images (acquired in near infrared and visible light) representing 259 cases, the research team engineered a software package, PMExpert, that incorporated three post-mortem specific iris matching algorithms. To understand what features humans believe to be useful in post-mortem iris matching, participants analyzed pairs of post-mortem samples, classified them as those originating from the same or different eyes, and annotated features supporting the decision.
Iris Images:
After the curation of all data collected by the Dutchess County Medical Examiner's Office, NY, iris images from 259 cases were selected for the final dataset release, and for analyses carried out in this project. This data corpus consists of 5,770 NIR and 4,643 RGB images, including images for one peri-mortem case with corresponding post-mortem samples after demise.
Human Examination Data:
The researchers conducted an experiment to collect annotation data on what humans believe to be distinctive features useful for post-mortem iris matching. Initial participants were recruited through the University of Notre Dame to complete study tasks in-person on-site. Due to the COVID-19 pandemic, the study design was later modified to be an online experiment recruiting participants through Amazon Mechanical Turk.
This data acquisition took place in two rounds:
- The first round was the initial collection of annotation data wherein participants had no prior knowledge of the task or previous decisions.
- The second round, called the verification step, is where the annotations collected in the first round were presented to future participants for them to either agree with or disagree with along with supporting annotations.
Software Package:
A software tool called PMExpert was created to provide a simple unified interface for all recognition methods, allowing them to be used in an operational setting.
PMExpert consists of two main components: a command line interface (CLI) and a graphical user interface (GUI). Both components are meant to allow examiners to use post-mortem iris recognition methods on images that are collected in their routine operations, offering not only similarity scores and decisions, but also additional information to equip examiners to make their final decision.
Citation View help for Citation
Export Citation:
Funding View help for Funding
Subject Terms View help for Subject Terms
Geographic Coverage View help for Geographic Coverage
Smallest Geographic Unit View help for Smallest Geographic Unit
None
Distributor(s) View help for Distributor(s)
Time Period(s) View help for Time Period(s)
Date of Collection View help for Date of Collection
Data Collection Notes View help for Data Collection Notes
-
Software included with this dataset is understood to be provided "AS IS". ICPSR MAKES NO REPRESENTATIONS AND EXTENDS NO WARRANTIES OF ANY KIND, EITHER EXPRESSED OR IMPLIED. THERE ARE NO EXPRESS OR IMPLED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, OR THAT THE USE OF THE SOFTWARE WILL NOT INFRINGE ANY PATENT, COPYRIGHT, TRADEMARK, OR OTHER PROPRIETARY RIGHTS. Notwithstanding, ICPSR, to the best of its knowledge and belief, has the right and authority to provide this software to Recipient.
-
Warning: Some of the materials in this study may be disturbing in nature. The materials include several close-up images of the eye/iris of deceased individuals.
-
The iris images and PI-developed software are available for download under DS0 Study-Level Files, Other. Please note the files are large and may take over an hour to download.
Study Purpose View help for Study Purpose
The purpose of this project is to deliver a complete methodology and software that enhance unidentified decedent systems with a capability for comparisons of ante-mortem and post-mortem iris images.
Study Design View help for Study Design
Iris image data acquisition: Iris images were collected from deceased individuals under the jurisdiction of the Dutchess County Medical Examiner's office. 259 cases were selected for use in this dataset. Near-infrared (NIR) images were acquired using the IriTech IriShield MK 2021U sensor. Visible-light (RGB) images were acquired using the OmniVision OV8865 color sensor embedded into a tablet.
Human examination data (round 1): For this initial phase, 130 participants were recruited from the University of Notre Dame for on-site study task completion. Participants were presented with 10 pairs of post-mortem iris images, a single pair at a time. They were asked to first identify if the images were from the same eye or different eyes, then to annotate at least five defining features on each eye (whether they matched or did not match) that led them to make their decision. Participants were also able to select a "Don't Know" option if they could not determine. The research team developed a website where participants could complete the study tasks.
Human examination data (round 2): Participants in this round were Amazon Mechanical Turk "Master" users who completed study tasks online. Similar to round 1, participants were presented with pairs of iris images and asked to identify if the images were from the same eye or different eyes, then to annotate matching or non-matching features to support their decision. Unique to round 2, participants were presented with the previous annotations from a user as well as the version without annotations.
Time Method View help for Time Method
Universe View help for Universe
- For post-mortem iris sample acquisition: deceased individuals arriving at the Dutchess County Medical Examiner's Office premises.
- For human examination study: adults between 18 and 65 years old.
Unit(s) of Observation View help for Unit(s) of Observation
Data Source View help for Data Source
Dutchess County Medical Examiner's Office, NY
Data Type(s) View help for Data Type(s)
Mode of Data Collection View help for Mode of Data Collection
Description of Variables View help for Description of Variables
- Iris image data: For each image file listed, variables include eye orientation (left/right; ante-mortem or post-mortem), post-mortem interval in hours, type of image, and session/capture numbers.
- Human examination data: For each task, variables include the participant's unique ID number, file names of images presented, post-mortem intervals for each image, true image classification (same eye/different eye), and time (seconds) spent examining an image pair. Collective results also include the decisions from round 1 and round 2, and the number of matching and non-matching features as annotated by participants.
Response Rates View help for Response Rates
Not applicable.
Presence of Common Scales View help for Presence of Common Scales
None.
HideOriginal Release Date View help for Original Release Date
2023-03-29
Version History View help for Version History
2023-03-29 ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major statistical software formats as well as standard codebooks to accompany the data. In addition to these procedures, ICPSR performed the following processing steps for this data collection:
- Checked for undocumented or out-of-range codes.
Notes
The public-use data files in this collection are available for access by the general public. Access does not require affiliation with an ICPSR member institution.
ICPSR usually offers files in multiple formats for researchers to be able to access data and documentation in formats that work well within their needs. If you have questions about the accessibility of materials distributed by ICPSR or require further assistance, please visit ICPSR’s Accessibility Center.