Anatol Maier

Anatol Maier, M. Sc.

Department of Computer Science
Chair of Computer Science 1 (IT Security Infrastructures)

Hi, I’m Anatol Maier and I received my M.Sc. degree in computer science from the Friedrich-Alexander University Erlangen-Nuernberg (FAU), Erlangen, Germany, in 2019. Then in November 2019, I joined the IT Security Infrastructures Lab at FAU as a Ph.D. student and am now part of the Multimedia Security Group. My research interests include reliable machine learning, deep probabilistic models, and computer vision with particular applications in image and video forensics.

Recent projects

Below, a brief overview and description of my previously conducted research can be found.

Toward Reliable Models for Authenticating Multimedia Content: Detecting Resampling Artifacts With Bayesian Neural Networks (IEEE ICIP 2020)

In multimedia forensics, learning-based methods provide state-of-the-art performance in determining the origin and authenticity of images
and videos. However, most existing methods are challenged by out-of-distribution data, i.e., with characteristics that are not covered in
the training set. This makes it difficult to know when to trust a model, particularly for practitioners with a limited technical background.
In this work, we make a first step toward redesigning forensic algorithms with a strong focus on reliability. To this end, we propose
to use Bayesian neural networks (BNN), which combine the power of deep neural networks with the rigorous probabilistic formulation
of a Bayesian framework. Instead of providing a point estimate like standard neural networks, BNNs provide distributions that express
both the estimate and also an uncertainty range. We demonstrate the usefulness of this framework on a classical forensic task: resampling detection.
The BNN yields state-of-the-art detection performance, plus excellent capabilities for detecting out-of-distribution samples. This is demonstrated for three pathologic issues in resampling detection, namely unseen resampling factors, unseen JPEG compression, and unseen resampling algorithms. We hope that this proposal spurs further research toward reliability in multimedia forensics.

Publications

2024

2023

2022

2021

2020