معرفی
Matej Vitek is an Assistant and researcher at the Faculty of Computer and Information Science, University of Ljubljana, where he is a core member of the Computer Vision Laboratory (CVL). His work centers on advancing biometric security through innovative computer vision techniques, with primary focus on lightweight sclera recognition systems.
His academic journey includes:
- BSc in Computer Science and Informatics and Mathematics and Physics (2015), University of Ljubljana
- MSc in Computer Science and Informatics and Mathematics and Physics (2018), University of Ljubljana
- PhD in Computer and Information Science (2024), University of Ljubljana
Vitek's research expertise spans computer vision, biometrics, and deep learning, with specialized focus on sclera recognition. His methodology emphasizes developing computationally efficient models suitable for mobile deployment while addressing critical challenges like model bias and segmentation accuracy. Past explorations include quantum computing circuits and game development, demonstrating interdisciplinary versatility. Current work integrates anomaly detection for deepfake identification and explainable AI frameworks for biometric systems.
Publication trends reveal consistent advancement in sclera biometrics through large-scale collaborative efforts, including benchmarking competitions and novel dataset creation. His work demonstrates progression from foundational segmentation studies toward optimized lightweight architectures and bias mitigation strategies, reflecting the field's evolution toward practical, ethical deployment.
Vitek actively contributes to two major ARRS-funded projects: J2-50065 'DeepFake DAD' (2023-2026) developing anomaly detection methods for deepfake identification, and J2-50069 'MIXBAI' (2023-2026) creating interpretable mechanisms for explainable biometric AI. Previously, he participated in the P2-0214 Computer Vision research program (2019-2024) and consulting initiatives.
As a CVL laboratory member, he collaborates within a specialized biometrics team led by Prof. Peter Peer (supervisor) and Prof. Vitomir Štruc (co-supervisor), working alongside researchers Peter Rot, Žiga Emeršič, and Blaž Meden. His technical environment combines academic research with practical implementation challenges in resource-constrained settings.
