Approved for: 2 GIAC CPEs
The face recognition (FR) community has been very successful in developing algorithms that deliver high performance for facial images taken under controlled circumstances (e.g., passport photos). While these developments have led to the wide-scale deployment of FR systems in applications for law enforcement and fraud prevention, a number of challenges remain in the field. This Face Recognition session will focus on two fundamental challenges: unconstrained face recognition and human adjudication of potential match pairs. This session will answer the question of where does the current state of the art stand in resolving the unconstrained face recognition issue and what sort of applications might this help address. The session will then pivot to address the question of determining how well do forensic examiners perform the task of adjudicating potential match pairs, and what steps are being taken to develop this growing field of forensic facial examination.
3:30-3:50
Face Recognition with Unconstrained Video
Terry Adams
3:50-4:10
DHS Non-Cooperative Biometrics Program
Patricia Wolfhope
4:10-4:30
Dataset Generation and Annotation for Unconstrained Facial Recognition
Rebecca Erikson
4:30-4:50
Characterizing the Face Recognition Performance Envelope
Patrick Grother
4:50-5:10
Fused Human Performance: A Benchmark for Algorithm Performance
Jonathon Phillips
5:10-5:30
FISWG & OSAC: Developing Standards for Forensic Facial Identification
Richard Vorder Bruegge