Talk: Deep Learning for Morphing Attacks - Generation and Detection in Biometrics

05:30 - 6:30 PM (IST), December 04, 2020
5th IAPR International Conference on Computer Vision & Image Processing

Morphing attacks have emerged as recent but efficient threats on face recognition systems. Given face recognition systems are widely used in civilian identity management (such as passport issuance, visa management), many operations will be impacted if morphing attacks are not detected and prevented. Based on the findings from EU-FP7-FIDELITY project, EU has started coordinating efforts to make the face recognition systems robust against morphing attacks. The talk will focus on deep learning approaches for morphing attacks from both spectrum – generation and detection.

Speaker's Biography

Kiran Raja obtained his PhD in computer Science from Norwegian University of Science and Technology (NTNU), Norway in 2016. He is faculty member at Dept. of Computer Science at NTNU, Norway. His main research interests include statistical pattern recognition, image processing, and machine learning with applications to biometrics, security and privacy protection. He was/is participating in EU projects SOTAMD, iMARS and other national projects. He has authored several papers in his field of interest and serves as a reviewer for number of journals and conferences. He is a member of EAB and chairs Academic Special Interest Group at EAB. He also works as a consultant for various national agencies within Norway.