The module Bildbasierte Computergrafik (Image-based Graphics) is a mandatory course for MI master students in the Visual Computing study path and a WPF for all other master computer science studies.
It is held during the WS.
A detailed description can be found in the HOPS system or here.
Topics covered are:
- Filtering Operations: Linear filters (Box-, Gauss-, Sinc-), non-linear filters (Median, Vector-Median), edge-preserving filters (Bilateral, Cross-Bilateral, Guided Image Filter), edge-detectors (Sobel, Laplace, Canny)
- Segmentation- and matting algorithms (Mean-Shift, k-means, Spectral Clustering, Rotoscoping, etc.)
- Feature detectors (Harris-Corner Detector, SIFT, SURF, etc.)
- Camera calibration (Projective geometry, camera models, lense distortion, calibration algorithms, like Bundle Adjustment) und tracking methods
- 3D reconstruction techniques (Feature Matching, Stereo Reconstruction, Multi-View Reconstruction, Structured Light, Shape-from-Shading, Shape-from-X, etc.)
- Warping and morphing for image interpolation
- Light fields
- Panorama and 360° panorama stitching
- Python, C++, (Cuda, OpenCL programming)
All content is provided through the ILIAS course Bildbasierte Computergrafik.
You can receive the password per e-mail request or during the lecture.
Exam and successful participation in the practicals