Skip to content
Technical University of Munich
Picture of Martin Knoche

Martin Knoche, M.Sc.

Technical University of Munich

Chair of Human-Machine Communication (Prof. Hemmert komm.)

Postal address

Arcisstr. 21
80333 München

Research Areas

  • Deep Learning
  • Computer Vision
  • Face Recognition single and multi camera
  • Person Identification single and multi camera
  • Super-Resolution single and multi camera

Publications

  • Knoche, Martin; Rigoll, Gerhard: Tackling Face Verification Edge Cases: In-Depth Analysis and Human-Machine Fusion Approach. , 2023 more… Full text ( DOI ) Full text (mediaTUM)
  • Knoche, Martin; Teepe, Torben; Hörmann, Stefan; Rigoll, Gerhard: Explainable Model-Agnostic Similarity and Confidence in Face Verification. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, 2023 more… Full text (mediaTUM)
  • Knoche, Martin; Elkadeem, Mohamed; Hörmann, Stefan; Rigoll, Gerhard: Octuplet Loss: Make Face Recognition Robust to Image Resolution. 2023 IEEE 17th International Conference on Automatic Face and Gesture Recognition (FG), IEEE, 2023 more… Full text ( DOI ) Full text (mediaTUM)
  • Knoche, Martin; Elkadeem, Mohamed; Hörmann, Stefan; Rigoll, Gerhard: Octuplet Loss: Make Face Recognition Robust to Image Resolution. , 2022 more… Full text ( DOI )
  • Hörmann, Stefan; Kong, Tianlin; Teepe, Torben; Herzog, Fabian; Knoche, Martin; Rigoll, Gerhard: Face Morphing: Fooling a Face Recognition System Is Simple! , 2022 more… Full text ( DOI ) Full text (mediaTUM)
  • Knoche, Martin; Hörmann, Stefan; Rigoll, Gerhard: Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness. Leibniz Transactions on Embedded Systems 8 (1), 2022, 01--1 more… Full text ( DOI ) Full text (mediaTUM)
  • Hörmann, Stefan; Xia, Zhibing; Knoche, Martin; Rigoll, Gerhard: A Coarse-to-Fine Dual Attention Network for Blind Face Completion. 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021), 2021 more… Full text ( DOI ) Full text (mediaTUM)
  • Knoche, Martin; Hörmann, Stefan; Rigoll, Gerhard: Cross-Quality LFW: A Database for Analyzing Cross- Resolution Image Face Recognition in Unconstrained Environments. 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021), 2021 more… Full text ( DOI ) Full text (mediaTUM)
  • Hörmann, Stefan; Cao, Zhenxiang; Knoche, Martin; Herzog, Fabian; Rigoll, Gerhard: Face Aggregation Network For Video Face Recognition. 2021 IEEE International Conference on Image Processing (ICIP), IEEE, 2021, 2973-2977 more… Full text ( DOI ) Full text (mediaTUM)
  • Hörmann, Stefan; Zhang, Zeyuan; Knoche, Martin; Teepe, Torben; Rigoll, Gerhard: Attention-Based Partial Face Recognition. 2021 IEEE International Conference on Image Processing (ICIP), IEEE, 2021, 2978-2982 more… Full text ( DOI ) Full text (mediaTUM)
  • Hörmann, S.; Knoche, M.; Rigoll, G.: A Multi-Task Comparator Framework for Kinship Verification. 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020) , 2020, pp. 863-867 more…
  • Hörmann, S.; Moiz, A.; Knoche, M.; Rigoll, G.: Attention Fusion for Audio-Visual Person Verification Using Multi-Scale Features. 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020), 2020, pp. 281-285 more…
  • Hörmann, S.; Knoche, M.; Babaee, Ma.; Köpüklü, O.; Rigoll, G.: Outlier-Robust Neural Aggregation Network for Video Face Identification. Proc. International Conference on Image Processing (ICIP 2019), 2019, pp. 1675-1679 more… Full text ( DOI )
  • Heyn, H.-M.; Knoche, M.; Zhang, Q.; Skjetne, R.: A System for Automated Vision-Based Sea-Ice Concentration Detection and Floe-Size Distribution Indication From an Icebreaker. Proc. 36th International Conference on Ocean, Offshore and Arctic Engineering, ASME 2017, 2017, V008T07A012 more… Full text ( DOI )
  • Knoche, M.; Merget, D.; Rigoll, G.: Improving Facial Landmark Detection via a Super-Resolution Inception Network. 39th German Conference on Pattern Recognition, GCPR 2017, Springer Lecture Notes in Computer Science, 2017 more… Full text (mediaTUM)

Projects

Data-driven utilization of Multimedia Content
Partners: ProSiebenSat.1 Media SE, munich media intelligence
Period: 01.03.2018 - 30.09.2018

Facial recognition in public areas consistent with fundamental rights
Partners: Uniscon GmbH, Tüv Süd Digital Services, AXIS GmbH
Period: 01.10.2018 - 30.09.2020

Teaching

Student projects

For inquiries regarding any student projects please apply using your most recent CV together with your grade report emphasizing your previous experience in this area and your desired starting date.

Open Topics

You can find all open topics here.

Finished Projects

2023

  • Machine Learning-based CI Test Results Analysis (Master’s Thesis)

2022

  • Human Performance on Cross-Resolution Face Verification (Bachelor´s Thesis)
  • Image resolution impact on activation maps (Research Internship)
  • Quintuplet Loss X-Resolution Face Recognition (Master’s Thesis)

2021

  • Generating and Evaluating a Dataset for Realistic Cross-Resolution Face-Recognition (Bachelor´s Thesis)
  • Single-Output-Neuron Cross-Resolution Face Verification (Master`s Thesis)

2020

  • Comparison of SOTA Super-Resolution-Methods in the domain of Face Recognition (Research Internship)
  • Implementation of "Visualization of Activations" in a state-of-the-art Face Recognition Network (Research Internship)

2019

  • Enhanced Deep Residual Networks for Single Image Super-Resolution (Scientific Seminar)
  • Single Image and Multi Image Super Resolution for Face Identification (Master´s Thesis)
  • Attention Mechanisms in an Appearance Matching Network (Research Internship)
  • Enhanced Deep Residual Networks for Single Image Super-Resolution (Scientific Seminar)
  • Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network (Scientific Seminar)

2018

  • Repulsion Loss - Detecting Pedestriants in a Crowd (Scientific Seminar)
  • Wide Compression - Tensor Ring Nets (Scientifc Seminar)
To top