Dr.-Ing. Marco Rudolph
Marco Rudolph left the Institut für Informationsverabeitung.
Publications and research activities from the time after the departure are not listed here.

Marco Rudolph studied computer science at the Leibniz University Hannover. In 2016 he was awarded by the university the 'Preis des Präsidiums' for his special academic archievements. He attained the master degree in October 2018. In his thesis 'Coding of Human Body Shapes using Neural Networks' he dealt with autoencoders on body meshes.

In November 2018 he started at the TNT to work towards his PhD. The main focus is on image-based anomaly detection for industrial defect detection. This is mostly based on statistical modeling of image features with normalizing flows. Based on his work in this area, he gave a keynote for the workshop "Industrial Machine Learning" at ICPR 2022.  As secondary research areas, he deals with human pose estimation and interpretable machine learning.

Mr. Rudolph is also active in teaching for the TNT for several years. In 2021, he was nominated by the Faculty of Electrical Engineering and Computer Science of the LUH for the "APFEL" award for excellent teaching for his supervision of the course "Machine Learning". Furthermore, he has supervised more than 10 final theses and student projects.

Show recent publications only
  • Conference Contributions
    • Mathis Kruse, Marco Rudolph, Dominik Woiwode, Bodo Rosenhahn
      SplatPose & Detect: Pose-Agnostic 3D Anomaly Detection
      Computer Vision and Pattern Recognition Workshops (CVPRW) , June 2024
    • Thomas Norrenbrock, Marco Rudolph, Bodo Rosenhahn
      Q-SENN: Quantized Self-Explaining Neural Networks
      AAAI Technical Track on Safe, Robust and Responsible AI, AAAI Press, Vol. 38, No. 19, pp. 21482--21491, Vancouver, Canada, February 2024, edited by Michael J. Wooldridge and Jennifer G. Dy and Sriraam Natarajan
    • Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn, Bastian Wandt
      Asymmetric Student-Teacher Networks for Industrial Anomaly Detection
      Winter Conference on Applications of Computer Vision (WACV), IEEE, Hawaii, USA, January 2023
    • Thomas Norrenbrock, Marco Rudolph, Bodo Rosenhahn
      Take 5: Interpretable Image Classification with a Handful of Features
      Progress and Challenges in Building Trustworthy Embodied AI @NeurIPS, December 2022
    • Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn, Bastian Wandt
      Fully Convolutional Cross-Scale-Flows for Image-based Defect Detection
      Winter Conference on Applications of Computer Vision (WACV), IEEE, Hawaii, USA, January 2022
    • Tom Wehrbein, Marco Rudolph, Bodo Rosenhahn, Bastian Wandt
      Probabilistic Monocular 3D Human Pose Estimation with Normalizing Flows
      International Conference on Computer Vision (ICCV), IEEE, October 2021
    • Bastian Wandt, Marco Rudolph, Petrissa Zell, Helge Rhodin, Bodo Rosenhahn
      CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild
      Computer Vision and Pattern Recognition (CVPR), IEEE, Virtual, June 2021
    • Marco Rudolph, Bastian Wandt, Bodo Rosenhahn
      Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows
      Winter Conference on Applications of Computer Vision (WACV), IEEE, Online, January 2021
    • Marco Rudolph, Bastian Wandt, Bodo Rosenhahn
      Structuring Autoencoders
      Third International Workshop on “Robust Subspace Learning and Applications in Computer Vision” (ICCV), August 2019
  • Journals
    • Jan Thieß Brockmann*, Marco Rudolph*, Bodo Rosenhahn, Bastian Wandt, (* equal contribution)
      The voraus-AD Dataset for Anomaly Detection in Robot Applications
      Transactions on Robotics, IEEE, Vol. 40, pp. 438-451, November 2023, edited by Wolfram Burgard
    • Lutz M. K. Krause, Emily Manderfeld, Patricia Gnutt, Louisa Vogler, Ann Wassick, Kailey Richard, Marco Rudolph, Kelli Z. Hunsucker, Geoffrey W. Swain, Bodo Rosenhahn, Axel Rosenhahn
      Semantic Segmentation for Fully Automated Macrofouling Analysis on Coatings after Field Exposure
      Biofouling, Taylor & Francis, Vol. 39, No. 1, p. 64–79, March 2023, edited by Len Evans
Other activities

Supervised Theses

Human Motion

  • Deep-Learning-based probabilistic 3D Human Pose Estimation (later accepted to ICCV 2021)
  • Deep-Learning-based Completion of Human Motion
  • Baby Motion Capture
  • Human Shape Estimation using Sparse Data

Anomaly Detection

  • Deep-Learning-based Anomaly Detection for Robot Applications using Machine Data
  • Feature Extraction for Deep-Learning-Based Anomaly Detection
  • Active Learning for Deep-Learning-based Anomaly Detection
  • Machine Learning for Optical Defect Detection in Industrial Production of Chassis Components

Others

  • Enhancement of Graph Neural Networks for Session-based Recommender Systems
  • Visualization and Analysis Tools for Normalizing Flows
  • Development of an Image-Based System for Object Detection with Semi-automatic Labeling of Training Data
  • Retrospective Estimation of Head Motion from Structural MRI