International Workshop on Heterogeneous Face Recognition (HFR)
co-located with

The international workshop on Heterogeneous Face Recognition (HFR) will be held in conjunction with IEEE FG 2017 in Washington DC on May 30, 2017. This workshop is among the first to solicit proposals and providing a platform to bringing together the face recognition community in specifically addressing the challenging research gap in the cross-modal face recognition. The workshop will bring together Government & industry representatives, law enforcement agencies and academics for this emerging area of research.

Topics of Interest

The topics of the workshop includes but are not limited to

  • Sensors ((visible, infrared, polarimetric, LiDAR, and emerging sensing technologies))

  • Face detection in the HFR scenario

  • Fiducial point detection / face alignment in HFR (in infrared, thermal, sketches)

  • Face recognition in the wild scenarios (CCTV etc)

  • Face recognition at a distance
  • Infrared (near-infrared, thermal) to visible face matching

  • Sketch (hand sketch, forensic sketch) to visible face recognition

  • Biometrics applications

  • Surveillance, Law enforcement applications


0900-0915 - Sign in, upload presentations

0915-0945 - Welcome & Overview of HFR

0945-1030 – Keynote by Dr. Chris Boehnen

Bio: Dr. Chris Boehnen is a Senior Program Manager at the Intelligence Advanced Research Projects Activity (IARPA) focused on biometrics, computer vision, and machine learning.  He is the PM for the Odin, Janus, N2N Challenge, and BEST programs.  He is also joint faculty at the University of Tennessee.  Dr. Boehnen was formerly the founder and team lead for the Secure Computer Vision Team at Oak Ridge National Laboratory (ORNL).  In his six years at ORNL he served as Principal Investigator on $11 million in funding spread over 24 different grants which he conceived, proposed, and successfully executed.  He received the ORNL Early Career Award for Engineering and 3 of his papers have received best paper awards at highly competitive conferences including best paper out of 133 submissions at BTAS 2016.  Dr. Boehnen received his B.S., M.S. and Ph.D. from the University of Notre Dame Computer Science and Engineering Department.  He has been a member of the biometrics research community since 2001 when he began working on the Face Recognition Grand Challenge. 

1030-1130 - Oral presentations

·         “Cross-modal facial attribute recognition with geometric features,” C. Bradley, J. Ventura, T.E., Boult

·         “Deep network shrinkage applied to cross-spectrum face recognition,” C. Reale, H. Lee, H. Kwon, R. Chellappa

·         “On matching visible to passive infrared face images using image synthesis & denoising,” N. Osia, T. Bourlai

1130-1230 - Panel Discussion: “Synergy of heterogeneous face recognition research between government, academia, and industry"

1230-1400 - Lunch Break

1400-1445- Keynote by Professor Guillermo Sapiro

Title: “Not Afraid of the Dark: NIR-VIS Face Recognition via Cross-spectral Hallucination and Low-rank Embedding”

Abstract: Surveillance cameras today often capture NIR (near infrared) images in low-light environments. However, most face datasets accessible for training and verification are only collected in the VIS (visible light) spectrum. It remains a challenging problem to match NIR to VIS face images due to the different light spectrum. Recently, breakthroughs have been made for VIS face recognition by applying deep learning on a huge amount of labeled VIS face samples. The same deep learning approach cannot be simply applied to NIR face recognition for two main reasons: First, much limited NIR face images are available for training compared to the VIS spectrum. Second, face galleries to be matched are mostly available only in the VIS spectrum. In this paper, we propose an approach to extend the deep learning breakthrough for VIS face recognition to the NIR spectrum, without retraining the underlying deep models that see only VIS faces. Our approach consists of two core components, cross-spectral hallucination and low-rank embedding, to optimize respectively input and output of a VIS deep model for cross-spectral face recognition. Cross-spectral hallucination produces VIS faces from NIR images through a deep learning approach. Low-rank embedding restores a low-rank structure for faces deep features across both NIR and VIS spectrum. We observe that it is often equally effective to perform hallucination to input NIR images or low-rank embedding to output deep features for a VIS deep model for cross-spectral recognition. When hallucination and low-rank embedding are deployed together, we observe significant further improvement; we obtain state-of-the-art accuracy on the CASIA NIR-VIS v2.0 benchmark, without the need at all to re-train the recognition system.

Bio: Guillermo Sapiro received his B.Sc. (summa cum laude), M.Sc., and Ph.D. from the Department of Electrical Engineering at the Technion, Israel Institute of Technology, in 1989, 1991, and 1993 respectively. After post-doctoral research at MIT, Dr. Sapiro became Member of Technical Staff at the research facilities of HP Labs in Palo Alto, California. He was with the Department of Electrical and Computer Engineering at the University of Minnesota, where he held the position of Distinguished McKnight University Professor and Vincentine Hermes-Luh Chair in Electrical and Computer Engineering. Currently he is the Edmund T. Pratt, Jr. School Professor with Duke University. G. Sapiro works on theory and applications in computer vision, computer graphics, medical imaging, image analysis, and machine learning. He has authored and co-authored over 400 papers in these areas and has written a book published by Cambridge University Press, January 2001. G. Sapiro was awarded the Gutwirth Scholarship for Special Excellence in Graduate Studies in 1991, the  Ollendorff Fellowship for Excellence in Vision and Image Understanding Work in 1992, the Rothschild Fellowship for Post-Doctoral Studies in 1993, the Office of Naval Research Young Investigator Award in 1998, the Presidential Early Career Awards for Scientist and Engineers (PECASE) in 1998, the National Science Foundation Career Award in 1999, and the National Security Science and Engineering Faculty Fellowship in 2010. He received the test of time award at ICCV 2011. G. Sapiro is a Fellow of IEEE and SIAM. G. Sapiro was the founding Editor-in-Chief of the SIAM Journal on Imaging Sciences.

1445-1530 – Invited talk by Professor Thirimachos Bourlai

Title: “Face Recognition under Challenging Conditions”

Abstract: Within the last two decades, we have noticed improvement in the performance of forensic face recognition (FR) systems in controlled conditions characterized by suitable lighting and favorable acquisition distances. However, over the years the technology has steadily progressed to tackling increasingly more realistic conditions rather than adequately handling only well-controlled imagery. Most related research emphasizes maintenance of high recognition performance while coping with increased levels of image variability. Among the most insidious problems of visible-spectrum based FR algorithms are (1) the variation in level and nature of illumination, (2) the fact that as the level of illumination decreases, the signal to noise ratio rises quickly, and thus automatic processing and recognition become very difficult, (3) dealing with degraded face images acquired at night and at long standoff distances etc. In order to address these issues, recent research has moved into the use of infrared (IR) imagery (e.g., intensified near infrared (NIR), Short Wave IR, Middle and Long Wave IR - thermal). This presentation will discuss MILab's research efforts in face recognition across the imaging spectrum, the importance of forensics operators, as well as the future of this technology

Bio: Thirimachos Bourlai is an assistant professor in the Lane Department of Computer Science and Engineering at WVU. He also serves as an adjunct assistant professor in the WVU School of Medicine, Department of Ophthalmology, and the Department of Forensic and Investigative Science. He is the founder and director of the Multi-Spectral Imagery Lab at WVU. After earning his Ph.D. in face recognition and completing a post-doctoral appointment at the University of Surrey (U.K.), Bourlai completed a second post-doc in a joint project between Methodist Hospital and the University of Houston, in the fields of thermal imaging and human-based computational physiology. He joined the staff at WVU in 2009 serving as a visiting professor and later as a research assistant professor in the Lane Department.

1530-1545 - Afternoon Break

1545-1645 - Oral presentations

·         “Lower resolution face recognition in surveillance systems using discriminant correlation analysis,” M. Haghighat, M. Abdel-Mottaleb

·         “Multi-level feature learning for face recognition under makeup changes,” Z. Zheng, C. Kambhamettu

·         “Extended spectral to visible comparison based on spectral band selection method for robust face recognition,” N. Vetrekar, R. Ramachandra, K. Raja, R. Gad, C. Busch

1645-1700 - Concluding remarks