A graduate of first-cycle studies in the field of Cybersecurity was recognized in this year's edition of the Engineer 4 Science 2024 competition in the „Cyber” category. Congratulations!

The bachelor's thesis by Izabela Majchrowska, entitled „Analysis of selected cancelable biometric systems in terms of template unlinkability and the impact on FAR and FRR coefficients" was recognized as the best dissertation submitted this year to the Cyber 4 Science 2024 competition.

Our student investigated biometric systems in which identity is confirmed by, for example, a fingerprint or iris scan. First, a biometric fingerprint template is created in the database, and during subsequent identity verifications the sample is compared with the one in the database.

– In my thesis, I investigated the iris of the eye. Only one template can be created from one sample, but there are algorithms for cancelable biometrics that protect us in the event that someone steals our template – explains the author of the awarded dissertation. – These algorithms enable additional processing, owing to which we can create many different identities. If someone steals one biometric template, we can delete it and create a completely new, independent one.

According to Izabela Majchrowska , we can generate an infinite number of such independent templates. This independence is called unlinkability. The protection consists in the fact that it is not possible to state that many identities belonging to one person and created from one biometric, actually belong to the same person.

– In my work, I examined two cancelable biometrics algorithms to check whether they are characterized by unlinkability and whether adding this additional processing positively or negatively influences the statistics, i.e. whether such a biometric system works better or worse – explains Izabela Majchrowska, who continues her studies at our Faculty, in the field of Artificial Intelligence. As she observes, biometry, which she has become very interested in, combines very well with artificial intelligence. She therefore decided that this is the ideal field where she will be able to combine technology, i.e. artificial intelligence, with what she likes to do, i.e. biometry.

Her bachelor's thesis is also highly appreciated by her supervisor, Dr. Wojciech Wodo from the Department of Fundamentals of Computer Science:

– First of all, this work was prepared in a very meticulous manner, and the task was difficult, because Iza Majchrowska decided to reproduce the results of certain scientific studies published in well-known sources, de facto not having full knowledge and documentation – says the supervisor of the awarded work. – The articles do not present in detail, step by step, how to perform a given study. They show certain concepts, often carried out with certain assumptions undisclosed in detail to the public.

The student took on the ambitious task of reproducing these studies and verifying the presented results, and whether the concepts they described can actually work in biometric systems.

– And that was a true challenge. What is most interesting, we managed to recreate and verify the studies, and the conclusions are very interesting. It turns out that to some extent the results were confirmed, but in several cases things did not work as described in the articles, which raises the question: why? – says Dr. Wojciech Wodo. “Is it a matter of the data set, or maybe too little information disclosed in the scientific article, or maybe the algorithm requires some correction. This type of information is also found in this work. To sum up, a very interesting, advanced work, representing the level of at least a good master's thesis and therefore deserves the award.

Let us add that the award in the competition organized by 4 Science Institute Sp. z o.o. is not the only distinction that Izabela Majchrowska received for her dissertation. The Wrocław branch of the Association of Polish Electrical Engineers awarded her with the second prize in the Artificial Intelligence category in the competition for the best diploma theses. Congratulations!