Face Morphing Detection: An Approach Based on Image Degradation Analysis
In 2014 a novel identity theft scheme targeting specific application scenarios in face biometrics was introduced. In this scheme, a so called face morph melts two or more face images of different persons into one image, which is visually similar to multiple real world persons. Based on this non authentic image, it is possible to apply for an image based identity document to be issued by a corresponding authority. Thus, multiple persons can use such a document to pass image based person verification scenarios with a single document containing an artificially weakened template. Currently there is no reliable existing security mechanism to detect this attack.
This paper introduces a novel detection approach for face morphing forgeries based on a continuous image degradation. This is considered relevant because the degradation approach creates multiple artificial self-references and measures the “distance” from these references to the input. A small distance (significantly smaller than the one to be expected from a pristine image) could be considered as an anomaly here, indicating media manipulations (e.g. caused by morphing). Our degradation process is based on JPEG compression with different compression values. The evaluation results of our detection approach are classification accuracies of 90.1% under laboratory conditions and 84.3% under real world conditions.
KeywordsDigital image forensics Detection of face morphing attacks
The work in this paper has been funded in part by the German Federal Ministry of Education and Research (BMBF) through the research programme ANANAS under the contract no. FKZ: 16KIS0509K. The author would like to thank Jana Dittmann and Andrey Makrushin for the initial ideas as well as the joint work with both of them and Christian Kraetzer for discussions on the approach evaluated in this paper.
- 1.Ferrera, M., Franco, A., Maltoni, D.: The magic passport. In: Proceedings of the IEEE IEEE International Conference on Biometrics, Clearwater, Florida, pp. 1–7 (2014)Google Scholar
- 2.Makrushin, A., Neubert, T., Dittmann, J.: Automatic generation and detection of visually faultless facial morphs. In: Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2017). VISAPP, vol. 6, pp. 39–50 (2017). ISBN: 978-989-758-227-1Google Scholar
- 3.Hildebrandt, M., Neubert, T., Makrushin, A., Dittmann, J.: Benchmarking face morphing forgery detection: application of StirTrace for impact simulation of different processing steps. In: Li, C.-T. (ed.) Proceedings of the International Workshop on Biometrics and Forensics (IWBF 2017), Coventry, UK, University of Warwick, 4–5 April 2017Google Scholar
- 5.Schetinger, V., Iuliani, M., Piva, A., Oliveira, M.: Digital Image Forensics vs. Image Composition: An Indirect Arms Race. CoRR abs/1601.03239 (2016)Google Scholar
- 6.Raghavendra, R., Raja, K., Busch, C.: Detecting morphed facial images. In: Proceedings of 8th IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS 2016), Niagra Falls, USA, 6–9 September 2016Google Scholar
- 8.Scherhag, U., Raghavendra, R., Raja, K.B., Gomez-Barrero, M., Rathgeb, C., Busch, C.: On the vulnerability of face recognition systems towards morphed face attacks. In: Li, C.-T.(ed.) Proceedings of the International Workshop on Biometrics and Forensics (IWBF 2017), Coventry, UK, University of Warwick, 4–5 April 2017Google Scholar
- 9.Gomez-Barrero, M., Rathgeb, C., Scherhag, U., Busch, C.: Is your biometric system robust to morphing attacks? In: Li, C.-T. (ed.) Proceedings of the International Workshop on Biometrics and Forensics (IWBF 2017), Coventry, UK, University of Warwick, 4–5 April 2017Google Scholar
- 10.Rosten, E., Drummond, T.: Fusing points and lines for high performance tracking. In: IEEE International Conference on Computer Vision, vol. 2, pp. 1508–1511 (2005). doi: 10.1109/ICCV.2005.104
- 11.Mair, E., Hager, G., Burschka, D., Suppa, M., Hirzinger, G.: Adaptive and Generic Corner Detection Based on the Accelerated Segment Test (2010). http://www6.in.tum.de/Main/Publications/Mair2010c.pdf. Access 7 Feb 2017
- 12.Shi, J., Tomasi, C.: Good features to track. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 593–600 (1994)Google Scholar
- 13.Low, D.: Object recognition from local scale-invariant features. In: Proceedings of the International Conference on Computer Vision (1999)Google Scholar
- 15.Wolf, A.: Portrait Quality (Reference Facial Images for MRTD). Version: 0.08 ICAO, Published by authority of the Secretary General (2017)Google Scholar
- 17.Landwehr, N., Hall, M., Frank, E.: Logistic model trees. In: Proceedings in Machine Learning, pp. 161–205 (2005)Google Scholar
- 18.Hancock, P.: Psychological image collection at stirling (pics) - 2d face sets - Utrecht ECVP. http://pics.psych.stir.ac.uk/. Accessed 21 Apr 2017
- 19.Luxand, Inc.: Luxand - detect and recognize faces and facial features with luxand facesdk (2016). https://www.luxand.com/facesdk/. Accessed 6 June 2017