The 2019 Biometric Technology Rally

Similar documents
Advances in Iris Recognition Interoperable Iris Recognition systems

Software Development Kit to Verify Quality Iris Images

Tools for Iris Recognition Engines. Martin George CEO Smart Sensors Limited (UK)

Title Goes Here Algorithms for Biometric Authentication

EF-45 Iris Recognition System

Roll versus Plain Prints: An Experimental Study Using the NIST SD 29 Database

Biometric Recognition: How Do I Know Who You Are?

Experiments with An Improved Iris Segmentation Algorithm

Responsible Data Use Assessment for Public Realm Sensing Pilot with Numina. Overview of the Pilot:

Global and Local Quality Measures for NIR Iris Video

Thoughts on Fingerprint Image Quality and Its Evaluation

Future Attribute Screening Technology (FAST) Demonstration Laboratory


Biometrics for Public Sector Applications

Workshop. Debbie Lilu, CTMA Director Bill Chenevert, Senior Program Manager Marc Sharp, Project Manager

Project Status Update

Biometrics for Public Sector Applications

MorphoTrust TM Iris Recognition

EFFICIENT ATTENDANCE MANAGEMENT SYSTEM USING FACE DETECTION AND RECOGNITION

3 Department of Computer science and Application, Kurukshetra University, Kurukshetra, India

PERFORMANCE TESTING EVALUATION REPORT OF RESULTS

WHITE PAPER Need for Gesture Recognition. April 2014

Biometrics. Duane M. Blackburn Federal Bureau of Investigation

S&T Stakeholders Conference - West January 14-17, 2008

Human Identification from Video: A Summary of Multimodal Approaches

MOBILITY RESEARCH NEEDS FROM THE GOVERNMENT PERSPECTIVE

An Un-awarely Collected Real World Face Database: The ISL-Door Face Database

ENHANCHED PALM PRINT IMAGES FOR PERSONAL ACCURATE IDENTIFICATION

Joint Industry Programme on E&P Sound and Marine Life - Phase III

Algorithm for Detection and Elimination of False Minutiae in Fingerprint Images

Distinguishing Identical Twins by Face Recognition

Iris Recognition using Histogram Analysis

S&T Stakeholders Conference

Second Symposium & Workshop on ICAO-Standard MRTDs, Biometrics and Security

ENHANCING THE USABILITY OF THE HUMAN MACHINE INTERFACE ON THE HANDHELD INTERAGENCY IDENTITY DETECTION EQUIPMENT (HIIDE)

On The Correlation of Image Size to System Accuracy in Automatic Fingerprint Identification Systems

CALL TO ARTISTS Cape Cod Healthcare s New Emergency Centers at Cape Cod Hospital and Falmouth Hospital

Challenges and Potential Research Areas In Biometrics

Biometrics - A Tool in Fraud Prevention

Job Description. Commitment: Must be available to work full-time hours, M-F for weeks beginning Summer of 2018.

City Research Online. Permanent City Research Online URL:

Eyedentify MMR SDK. Technical sheet. Version Eyedea Recognition, s.r.o.

Loyola University Maryland Provisional Policies and Procedures for Intellectual Property, Copyrights, and Patents

Student Attendance Monitoring System Via Face Detection and Recognition System

Rank 50 Search Results Against a Gallery of 10,660 People

E90 Project Proposal. 6 December 2006 Paul Azunre Thomas Murray David Wright

ISO/IEC JTC 1/WG 11 N 49

Earth Station Coordination/Notification and Tools

SIMULATION-BASED MODEL CONTROL USING STATIC HAND GESTURES IN MATLAB

International Journal of Scientific & Engineering Research, Volume 7, Issue 12, December ISSN IJSER

TRUESENSE SPARSE COLOR FILTER PATTERN OVERVIEW SEPTEMBER 30, 2013 APPLICATION NOTE REVISION 1.0

CymbIoT Visual Analytics

Development and Integration of Artificial Intelligence Technologies for Innovation Acceleration

Evaluation of Biometric Systems. Christophe Rosenberger

Research Seminar. Stefano CARRINO fr.ch

Chapter 6 Face Recognition at a Distance: System Issues

FEATURES Industry windows paperless solutions High speed portable document scanner is well-suited for a wide variety of Window industry

Notification and recording of frequency assignments in Space Services

Vein and Fingerprint Identification Multi Biometric System: A Novel Approach

INTERNATIONAL RESEARCH JOURNAL IN ADVANCED ENGINEERING AND TECHNOLOGY (IRJAET)

OPPORTUNISTIC TRAFFIC SENSING USING EXISTING VIDEO SOURCES (PHASE II)

Department of Computer Science & Engineering Michigan State University December 10, 2010

REPORT FROM THE COMMISSION TO THE EUROPEAN PARLIAMENT AND THE COUNCIL. on the evaluation of Europeana and the way forward. {SWD(2018) 398 final}

Outdoor Face Recognition Using Enhanced Near Infrared Imaging

Pan-Canadian Trust Framework Overview

The Use of Static Biometric Signature Data from Public Service Forms

ANALYSIS OF PARTIAL IRIS RECOGNITION

Multimodal Face Recognition using Hybrid Correlation Filters

Hand Segmentation for Hand Gesture Recognition

FACE VERIFICATION SYSTEM IN MOBILE DEVICES BY USING COGNITIVE SERVICES

D8.1 PROJECT PRESENTATION

Background. Computer Vision & Digital Image Processing. Improved Bartlane transmitted image. Example Bartlane transmitted image

Empirical Evidence for Correct Iris Match Score Degradation with Increased Time-Lapse between Gallery and Probe Matches

The Hong Kong Institute of Directors Directors Conference 2017

Notification and Recording of Frequency Assignments in Space Services

IMAGE ENHANCEMENT. Quality portraits for identification documents.

CAPACITIES FOR TECHNOLOGY TRANSFER

Automated Planetary Terrain Mapping of Mars Using Image Pattern Recognition

COMMISSION RECOMMENDATION. of on access to and preservation of scientific information. {SWD(2012) 221 final} {SWD(2012) 222 final}

move move us Newsletter 2014 Content MoveUs has successfully finished the first year of the project!

Near Infrared Face Image Quality Assessment System of Video Sequences

Applied Surveillance using Biometrics on Agents Infrastructures

Feature Extraction Techniques for Dorsal Hand Vein Pattern

the role of mobile computing in daily life

Command, Control and Interoperability

ARCHIVED. Disclaimer: Redistribution Policy:

Facial Recognition of Identical Twins

The Disappearing Computer. Information Document, IST Call for proposals, February 2000.

Recent research results in iris biometrics

Controlling vehicle functions with natural body language

Little Fingers. Big Challenges.

Facial Biometric For Performance. Best Practice Guide

ABSTRACT INTRODUCTION. Technical University, LATVIA 2 Head of the Division of Software Engineering, Riga Technical University, LATVIA

International Journal of Advanced Research in Computer Science and Software Engineering

Foundations Required for Novel Compute (FRANC) BAA Frequently Asked Questions (FAQ) Updated: October 24, 2017

INTERNATIONAL JOURNAL OF PURE AND APPLIED RESEARCH IN ENGINEERING AND TECHNOLOGY

1. Redistributions of documents, or parts of documents, must retain the SWGIT cover page containing the disclaimer.

Operational Objectives Outcomes Indicators

THE GALLERY AT BAYSIDE ARTS & CULTURAL CENTRE BOARD STRATEGIC PLAN

NFC Forum NFC Use Cases

Transcription:

DHS SCIENCE AND TECHNOLOGY The 2019 Biometric Technology Rally Kickoff Webinar, November 5, 2018 Arun Vemury -- DHS S&T Jake Hasselgren, John Howard, and Yevgeniy Sirotin -- The Maryland Test Facility Biometrics & Identity Technology Center Department of Homeland Security Science and Technology Directorate

Biometrics & Identity Technology Center Vision Provide S&T s core biometric RDT&E capability to drive enduring efficiencies and biometric & identity innovations across the HSE Facilitate accelerated adoption for emerging DHS use cases Follow Build one, use widely approach Goals Support cross cutting methods, best practices, and solutions across programs to drive efficiencies Provide objective biometric & identity test and evaluation services to DHS Engage Industry and provide better feedback Encourage Innovation across the HSE Industry Collaboration Testing & Evaluation Address Growing Needs DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 2

Biometrics Technology Rallies Challenge industry to meet specific DHS use-cases Foster innovation and create partnerships across government and industry Inform DHS procurement activities such as operational tests, pilots, and system acquisitions Participation provides industry: An opportunity to collect and retain biometric data for research purposes Obtain user feedback Receive a performance report Showcase system performance to potential customers The use-case for the 2018 Biometric Technology Rally was to identify users in a high-throughput security environment with unmanned face/iris systems. DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 3

Goals for the 2019 Biometric Technology Rally Target the high-throughput use case Face, fingerprint or iris imagery Acquisition systems and/or matching algorithms Measure progress from 2018 Rally in regards to efficiency effectiveness and satisfaction Measure (for the first time) high-throughput fingerprint systems Access the ability of acquisition systems to capture images that work across algorithms Access the ability of algorithms to work across acquisition systems Collaborate and guide promising technologies, share information via Cooperative Research and Development Agreements DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 4

2019 Biometric Technology Rally System Requirements Who will be able to participate: Vendors of face, iris, fingerprint or multi-modal biometric acquisition systems Vendors of face, iris, fingerprint, or multi-modal biometric matching algorithms Minimum requirements for acquisition systems: Operate in an unmanned mode, i.e. without an operator/instructor Operate within the physical footprint and infrastructure defined by the MdTF (6 x8 ) Collect biometric imagery to support identification operation Provide at least one biometric image per Test Volunteer Process and submit biometric data within defined time constraints (TBD still H.T.) Optional Provide up to three images per modality per Test Volunteer Minimum requirements for matching algorithms: Provide a software library that confirms to the MdTF Simplified Biometric Algorithm Interface (MSBIA) Operate within some computational limits (RAM/CPU usage, speed, etc. TBD) DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 5

Acquisition Systems Systems may submit face, iris, or fingerprint* samples or some combination thereof (multimodal) Can submit up to three samples per type (face, left iris, left little finger, etc.) per subject Systems are no longer required to submit face image(s)* * Indicates a change between 2018 Rally and 2019 Rally DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 6

Acquisition Systems Evaluation Metrics Efficiency (overall) Refers to the amount of time required to use each biometric system Quantified as average transaction time (beam-break to beambreak) for Test Volunteers at each Rally System Effectiveness (per modality) Refers to the accuracy and completeness with which users are identified. Measured in two time intervals: By 5 seconds after the entry beam break By 20 seconds after the entry beam break Failure to Acquire Rate (FtAR) for images Proportion of Test Volunteers for whom no images were captured True Identification Rate (TIR) for images The proportion of Test Volunteers correctly identified MdTF matcher (for 2018 Rally comparison) and other matchers Satisfaction (overall) Refers to Test Volunteers positive attitudes toward the Rally Systems Measured using a 4-button kiosk from Very Happy to Very Unhappy Quantified as proportion of Happy or Very Happy responses Efficiency Effectiveness: FtAR TIR Satisfaction DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD

Acquisition System Evaluation Metrics Systems are expected to meet minimum requirements: Identify at least 95% of all volunteers Maintain average transaction time of 10 seconds Maximum transaction time of 20 seconds Data collection will be stopped if allotted time is exceeded Average top-box satisfaction score of > 90% Recognize when quality sample has been collected Adjust when better samples are needed. Given multiple attempts to capture a sample, Systems should: Have a decreasing number of overall captures per opportunity (recognize) Have increasing relative quality in latter capture opportunities (adjust) Performance will be assessed based on the last image sent DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 8

Matching Algorithms Provides a new option to apply to participate in the Rally Selected face, finger, or iris algorithms will need to conform to a simplified common interface We will use your algorithm to match biometric samples from 2018 and 2019 Rallies Matching algorithm vendors are not required to provide a biometric acquisition system Biometrics acquisition system vendors are not required to provide a matching algorithm DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 9

Matching Algorithm Evaluation Metrics Each acquisition system will collect images from 300-400 test subjects During the test, test volunteers will participate in a manned enrollment to create a same-day gallery of face, finger, and iris samples Most test volunteers have previously provided face, finger, and iris samples which are part of the historic gallery Each algorithm will be used to compare samples from the acquisition systems to the same-day and historic galleries Results will show: Overall performance Robustness to collection conditions Robustness to longitudinal effects System 1 System 2 System 3 finger face face iris Same-day Gallery Historic Gallery Finger Matcher A Face Matcher B Iris Matcher B DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 10

TIR TIR Acquisition and Algorithms What is the additional challenge? Acquisition system: High match rate with your acquired images across matching algorithms Robustness to matching algorithm Matching algorithm: High match rate from your algorithm across all acquisition systems Robustness to capture conditions 11

Why participate in the 2019 Rally? For Acquisition System Participants Acquisition system participants will get statistics and videos of error cases to help improve their products Acquisition system participants will get samples collected by their system on a diverse population of 300-400 people 82% of Rally Participants learned something about their systems during the Rally and 64% plan to make improvements so that their systems are better suited to DHS use-cases For Matching Algorithm Participants Algorithm participants will get enrollment images and metadata for the diverse rally volunteers Algorithm participants will receive results of their algorithm per rally volunteer, per Rally acquisition system to enable better understanding of variation in algorithm performance Opportunities to partner with acquisition system participants DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 12

Why participate in the 2019 Rally? Industry Recognition Aliased Reporting Over 2,250 visitors to mdtf.org/rally since November 2017 72 different countries, 40% of traffic is non-us Visitors spent on average 3 minutes digesting results Linked to by dozens of press releases, news articles, social media, etc. Showcase systems via VIP day On-going Cooperative Research and Development Agreement with DHS S&T DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 13

Acquisition System Application Package Submission deadline: 11/30/2018 Provide an application package (limit 5 pages), in the form of a white paper, that addresses the following: 1. Overview of the acquisition system a) Description of modalities acquired, cameras or sensors utilized, and other equipment necessary (computers, flow management hardware, etc.) b) Expected layout of all equipment within constrained space (6 x8 ) c) Network and power requirements d) Example use case and workflow 2. Descriptions of the complexity and maturity of the acquisition system a) When was the system first conceived and developed? Is it still under development? b) Known acquisition issues (height, weight, or disability restrictions, etc.). c) Has a third party integrated your system into a larger system? If so, how much effort was needed? d) Describe any past or present operational deployments. DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 14

Acquisition System Application Package 3. Descriptions of the imagery acquired by the system a) Sample types produced (slap vs four single print for example) b) Formats (jpg, png, wsq, etc.) c) Quantity and Quality 4. Description of user interaction with the system a) Actions the users perform to complete a transaction b) Instructions / feedback provided to users c) Are there any exception processes? 5. Estimates of Performance a) Estimated failure to acquire rate b) Estimated true positive identification rate using a top 3 recognition algorithm c) Estimated transaction time DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 15

Acquisition System Application Package 6. System safety information, including eye safety a) Demonstrate why system is safe for human users b) Does your system use specific wavelengths for illumination? c) Are there any sources of exposed current? d) Are there any exposed sharp edges or moving parts that could cause physical harm? 7. Demonstration video of system functionality a) Provide up to 2 minutes of video demonstrating your system in use. b) Should reflect information included in white paper. c) If video is too large to include in an email, provide a URL for reviewers to access. DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 16

Matching Algorithm Application Package Submission deadline: 11/30/2018 Provide an application package (limit 5 pages), in the form of a white paper, that addresses the following: 1. Overview of the matching algorithm a) Modalities and acceptable biometric sample types (slap vs. individual finger for example) b) High level overview of the underlying technology (CNN, Gabor wavelets, Haar cascades, etc.) c) Recommended CPU*, RAM, disk, operating system and runtime dependencies. d) What programming languages does the algorithm SDK support? 2. Descriptions of the complexity and maturity of the matching algorithm a) When was the algorithm first conceived and developed? Is it still under development? b) Known processing issues (image size, pixels between the eyes, occlusion, pose, or gaze angle restrictions, etc.). * Algorithms that require GPUs for execution are out of scope for 2019 Rally. DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 17

Matching Algorithm Application Package 2. Descriptions of the complexity and maturity of the matching algorithm c) Has a third party integrated your algorithm into a larger system? If so, how much effort was needed? d) Has the algorithm been a part of any formal, public testing, such as NIST FRVT? e) Describe any past or present operational deployments. 3. Estimates of Performance a) Failure to Process Rate (unable to create a template) b) True Match Rates and matching thresholds for the following False Match Rates: i. @ 1:10,000 FMR. What is the match threshold and the expected TMR? ii. iii. @ 1:100,000 FMR. What is the match threshold and the expected TMR? @ 1:1,000,000 FMR. What is the match threshold and the expected TMR? c) Estimates of algorithm stability to pose, motion blur, low contrast and images from a diverse range of acquisition systems DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 18

What to Expect (Acquisition and Matching Systems) Call for Participation Application Deadline Conditional Acceptance Cloud API Finalized 11/5/2018 11/30/2018 1/4/2019 01/21/2019 Acceptance Notification Installation at the MdTF Stakeholders VIP Day Rally Test 02/18/2019 04/01/2019 04/03/2019 04/05/2019 6 days of collection DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 19

What to Expect Upon submitting an application package, you will be notified on whether your system has been accepted for participation in the 2019 Biometric Technology Rally approximately 5 weeks after the deadline. An accept into the Biometric Technology Rally will be considered conditional until A Cooperative Research and Development Agreement (CRADA) has been signed by both your organization and DHS S&T. You have demonstrated proper integration into the MdTF API. You have provided the proper safety certifications/documents. Members of your team who will be present at the MdTF during testing have: completed the proper human subject testing training. completed any MdTF specific training. DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD 20

Thank You! Questions? peoplescreening@hq.dhs.gov For more information: peoplescreening@hq.dhs.gov http://mdtf.org All 2018 Biometric Technology Rally Results Material from all webinars, briefings, outreach Announcements about future Rally s DHS Science and Technology Directorate MOBILIZING INNOVATION FOR A SECURE WORLD