After successful completion of the course, students are able to solve first problems in the fields of in machine vision: basic computer vision methods, edge detection, region description, feature extraction, object tracking, depth image acquisition, methods of 2D and 3D object recognition, Deep learning and Machine Learning, Gestalt theory, depth image processing, cognitive vision; Focus in robotics on cognitive robots, situated vision for robotics, and robot systems.

Emphasis is on the following topics in machine vision: basic computer vision methods, edge detection, region description, feature extraction, object tracking, depth image acquisition, methods of 2D and 3D object recognition, Gestalt theory, depth image processing, cognitive vision; Focus in robotics on cognitive robots, situated vision for robotics, and robot systems.

• Robots, robot tasks, cognitive robots, machine vision, vision applications; computer/machine/situated vision, and machine vision basics: camera, images, Filtering, SSD, Canny
• Machine Vision Features: Industrial/mobile/cognitive robotics, sensors used in robotics;
• Interest Points: Harris, DoG
• Object recognition 2D: SIFT, SURF
• Geometry Stereo: geometry, basic calibration, stereo vision 
• 3D Camera Systems: Other methods to obtain 3D images
• Attention
• Ransac
• 3D Vision Methods: voxel grids, neighbours, integral images, surface normal, differential geometry, Gestalt, Clustering
• Object recognition in 3D: NARF, VFH, ESF, ..., examples, learning from CAD data
• Deep learning: concept, introduction, applications, object categorisation
• Open problems: human vision vs. robot vision, what works and open challenges

The contents of the course are presented in the form of oral presentations supported by digital slides. To apply the theoretic knowledge, students have to implement 6 programming exercises about topics discussed during the lecture. The exercises have to be done using Python with the popular libraries OpenCV, Open3D and NumPy. The submission and grading of the exercises is done via TUWEL.

Introduction and first lecture: 2. October 15:00-18:00

Positive of all exercises followed by oral examination. Weight for final grade: Exercises 60%, Oral Examination 40%.

Due to the current situation, exams are offered both in attendance and ALSO online with Zoom. Whether an appointment is an attendance exam or an online exam can be found in the information in TISS under the individual exam dates.

Examination candidates will be assigned a time slot of 30 minutes. This slot will be sent out by mail after the end of the registration deadline. A webcam is needed for proof of identity. The public link of the exam will be shared in TISS. Candidates with time constraints or restrictions in terms of equipment should contact us before the exam. 

Knowledge of Python is recommended. Background in Robotics is helpful, e.g., 376.040 Fachvertiefung Bildverarbeitung und Robotik.