• Finished Master Theses

      • Lars Haalck: Approaching Critical Configurations in Monocular Structure-from-Motion for In-Field Animal Tracking
      • Frederik Probst: CNN-based Animal Classification in a Novel Camera Trap Database
      • Christian Lentfort: Laser Light Plane Based Motion Analysis: Construction of a Prototype
      • Christof Duhme: Deep learning-based Blood Vessel Segmentation in Neurosurgical Angiography Recordings
      • Thomas Teodorowicz: Random Forest-based Active Learning for Angiographic Image Segmentation
      • Christian Wollny: Uncertainty Sample-based Active Learning Through Associative Reinforcement Learning
      • Julian Hesse: Learning to See in the Noise - Deep Learning-based Image Restoration of Short Exposure EM Images
      • Karim Huesmann: Real-time Layer-wise Analysis of Convolutional Neural Networks
      • Sebastian Thiele: Capsules Only Look Once - Comparing State-of-the-Art Deep Learning Architectures in a Novel Coats of Arms Image Domain
      • Christoph Blecke: CamGanAPI: A Modular Camera Gantry Programming Interface for Reactive Additive Manufacturing
      • Dominik Berse: Machine Learning for Additive Manufacturing: Analysing Multiple Accelerometer Readings for 3D Printing
      • Eike Gebauer: Real-time Monitoring of 3D Printers: Machine Learning Based Anomaly Detection by Monitoring the Load of Multiple Stepper Motors
      • Mareen Hoffmann: Automatic Analysis of Sperm Mobility for Reproductive Medicine - Comparing a Deterministic and a Probabilistic Tracking Approach for Small Objects Under the Consideration of Drift
      • Marvin Stuffert: Possibilities and Limitations of a Real-Time Insect Camera Trap
      • Pascal Kockwelp: Machine Learning Meets Polarisation: Approximating Polarised Light Properties based on Static RGB Images using Fully Convolutional Neural Networks
      • Bafrin Krad: Analysis of Age-related Macular Degeneration in OCT Images with Deep Convolutional Neural Networks
      • Daniel Müller: Considerate Additive Manufacturing: A Machine Learning-based Closed-Loop Skin Detection and Collision Avoidance System for 3D Printing
      • Hans-Christian Lindner: Developing a robotic lawn mower using a convolutional neural network for object detection with an external camera and different SLAM algorithms for navigation
      • André Machate: Evaluation of deterministic and probabilistic deep learning models to improve short exposed electron microscope images
      • Julian Bigge: Hatching-Box: Monitoring the Rearing Process of Drosophila Using an Embedded Imaging and in-Vial Detection System
      • Leonard Allewelt: Automatic analysis of sperm movement and roll behavior using Deep Neural Networks
      • Jacqueline Kockwelp: Deep Learning-based Analysis of High Resolution Cytological Imagery for AML Prognosis
      • Jan Ewald: Machine learning-based forensic analysis of images and videos for deepfake detection

       

Finished Bachelor Theses

  • Viktor Gorte: Comparing Closed and Open Source SfM Pipelines to Reconstruct Natural Environments
  • Raoul Kanschat: Efficient Key-Frame selection and Panorama Stitching for Geospatial Visual Animal Tracking
  • Jan Phillip Bläs: Entropy as a Measure to Quantify the Training Quality of Neural Networks
  • Gabriela Cristina Feldhaus: Quantifying the Cyclic ICG-Flow in Angiography Videos
  • Clemens Kohl: Resolving Collisions of Visually Indistinguishable Objects using Fully Convolutional Networks
  • Allan Grunert: Optical Flow-based Detection of Small Objects in Cluttered Environments
  • Jacomo Axel Krause: Advanced Support and Infill Generation for Additive Manufacturing
  • Jacqueline Kockwelp: Evaluating Different Deep Learning Architectures To Resolve Colliding Drosophila Melanogaster Larvae
  • Timm Jasper Kühnel: An Empirical Study of Critical Configurations in Monocular Structure-from-Motion
  • Anatoli Dick: Compensating Unbalanced Data using Style Transfer for Machine Learning-based Camera Trap Image Classification
  • Julitta Sucker: Towards Parameter-free Deep Learning in Biomedical Applications - Evaluating Multiple Early Stopping Criteria for Image Segmentation