6th IAPR/IEEE International Workshop on Biometrics and Forensics
June 7, 8 2018, Sassari, IT
Professor Mark Nixon
School of Electronics and Computer Science University of Southampton, UK
Title: Biometrics, Forensics and Identity Science
The main question I shall address is whether biometrics can be used in forensics. There have been a number of studies in this area already and my main focus is on biometrics that can be used for identification in surveillance video. There are now papers which show that the human body can offer better identification performance than the face. This is not reflected in current practice and direction. I shall first introduce my view of biometrics and forensic identification. And then give case studies on gait and ear identification. I shall also describe using semantics for identification before conclusions and recommendations for future work. I contend that biometrics can indeed be used in forensics, but that development should be directed by biometric performance rather than by notions of human perception.
Mark S. Nixon is currently a Professor of computer vision with the University of Southampton, U.K. His research interests are in image processing and computer vision. His team develops new techniques for static and moving shape extraction, which have found application in automatic face and automatic gait recognition and in medical image analysis. His team were early workers in face recognition, later came to pioneer gait recognition, and then joined the pioneers of ear biometrics. Recently his team have been pioneering soft biometrics for human identification. He has supervised many research projects with funding from sources, including DARPA, U.S. Army, EPSRC, NERC, EU, and General Dynamics. His vision textbook, with A. Aguado, Feature Extraction and Image Processing, Third Edition (Academic Press, 2012) has become a standard text in computer vision. With T. Tan and R. Chellappa, their book Human ID Based on Gait is part of the Springer Series on Biometrics and was published in 2005. He is the President of the IEEE Biometrics Council. He is a member of IAPR TC4 Biometrics. He is a fellow of the IET and the IAPR, and a Distinguished Fellow of the BMVA. He has chaired/program chaired many conferences, such as BMVC, ICPR, IEEE BTAS, IAPR ICB, IAPR/IEEE IJCB, IEEE ISBA, and IAPR/IEEE IJCB 2017 and given many invited talks.
Dr Giovanni Tessitore
Director of Electronic and Digital Evidence Section at Forensic Science Police Service of Polizia di Stato, Italy
Title: Face Analysis for Forensic Cases
Forensic Facial Image Comparison (FIC) is currently a subjective evaluation made by a forensic specialist in order to assess the correspondence between two given facial images, one of a known person (suspected) and the other of an unknown person. The forensic specialist should subjectively and empirically estimates the strength-of-evidence of two opposite hypothesis (on the basis of the observed similarities/dissimilarity): the accusatory hypothesis – the two photos belong to the same person; the defensive hypothesis – the two photos belong to different people. Despite a lot of work has been recently done in order to standardize the methodology applied by experts (morphological approach with the use of a checklist of features) FIC still remains a subjective evaluation. Automatic Facial Recognition Systems (AFRS) could be used, in principle, to obtain an objective evaluation instead. However scores obtained with AFRS have no meaning in a forensic scenario where results should always be expressed in terms of strength-of-evidence. At such purpose the problem of estimation of likelihood-ratio for the scores resulting from AFRS in the Bayesian Framework will be introduced together with the related evaluation procedure, in which standard ROC or DET curves are replaced by the log-likelihood-ratio cost (Cllr).
Giovanni Tessitore has a degree in Computer Science with a PhD in Computational and Computer Science at the University of Naples Federico II. During the research, activity carried out during three years of doctorate and in the following two years of research as Post Doc mainly dealt with statistical learning methods (Neural Networks, Support Vector Machines, etc.) applied to the modelling of neural processing of vision and object/action recognition. Since 2012, he has been working for the Forensic Science Police Service of Polizia di Stato. He currently directs the Electronic Investigation Section of the IV Division.