Slides will be posted as the course pregresses.
- Topic 1. Introduction
- Topic 2. Overview of different image modalities: photo images,
video, and 2D-3D-4D medical data
- Topic 3. Elements of image (pre)-processing and feature detection: point processing, filtering, gradients, colors, patches, edges, corners, etc.
- Topic 4. Basic image segmentation and general clustering methods (thresholding, region-growing, K-means, Normalized Cut, mean-shift)
- Topic 5. Boundary regularization: livewire, deformable models (snakes), gradient descend, DP-snakes.
- Topic 6. Combining color and boundary. Binary graph cuts. Fixed appearance models. Color model fitting (Gaussians, histograms) and clustering (variance, entropy). Set functions and optimization. Submodularity. Non-submodular and higher-order energies.
- Topic 7. Correspondence. Stereo. Local (windows), scan-line (DP), and global (graph cuts) approaches to stereo.
- Topic 8. Multi-label image analysis. Convex and robust pairwise potentials. Optimization (ICM, simulated annealing, Ishikawa, multi-way graph cuts, a-expansions). Regularization models (MRF, MDL, sparsity). Applications (restoration, stitching, detection, motion, segmentation).
- Topic 9. Geometric models. RANSAC. Multi-model fitting. Some structure and motion problems.