Fundamentals of imaging / Principles of machine learning for computer vision, model families, and training/regularization concepts / Data and experiment pipelines: acquisition, annotation, versioning, valid splits; metrics and acceptance criteria / Learning with scarce data: transfer learning, semi-/self-supervised learning, anomaly detection, few-/active learning, synthetic data & augmentation / Industrial computer-vision tasks (classification, detection, segmentation, tracking, OCR/pose estimation) / Robustness: environmental/operational perturbations, stress testing, domain/data shift, out-of-distribution detection / Transparency: uncertainty quantification (incl. calibration) and explainable artificial intelligence / Deployment & integration: edge/cloud, model export, optimization, latency/cycle-time; monitoring, drift detection, and MLOps basics.

The students

• know industrial use cases and core computer-vision tasks, fundamentals of imaging, and the relevant metrics and requirements; they also gain an under-standing of typical industrial challenges such as data and label scarcity and learn methods to address these,
• learn principles of machine learning-based computer vision, including neural network architectures and training concepts, basic optimizations, monitoring and drift detection, as well as robustness-enhancing techniques such as data augmentation,
• can build data and experiment pipelines, adapt and train pretrained deep-learning models for industrial tasks, and evaluate them with appropriate metrics, including alignment with acceptance criteria.

• Linear algebra and calculus at the level of a bachelor’s degree in electrical engineering or information technology
• Basic probability and statistics at the level of a bachelor’s degree in electrical engineering or information technology
• Introductory programming skills (e.g., Python, MATLAB, or C++).
• Basic understanding of machine learning concepts.

• Szeliski, Richard: Computer Vision: Algorithms and Applications (2nd edition, online version available).
• Bishop, Christopher: Deep Learning: Foundations and Concepts (online version available).
• Forsyth, David; Ponce, Jean: Computer Vision – A Modern Approach (2nd edition).
• Goodfellow, Ian; Bengio, Yoshua; Courville, Aaron: Deep Learning (online version available).

Students are further encouraged to familiarize themselves with reading scientific papers early on, as these will be discussed and referenced throughout the course.

• Oral examinations take respectively place at the end of the summer semester and the winter semester.
• Further information and the exact dates can be found under the subheading Prüfungen »