The course studies image processing, image understanding, and video sequence analysis. Image processing deals with deterministic and stochastic image digitization, enhancement. restoration, and reconstruction. This includes image representation, image sampling, image quantization, image transforms (e.g., DFT, DCT, Karhunen-Loeve), stochastic image models (Gauss fields, Markov random fields, AR, ARMA) and histogram modeling. Image understanding covers image multiresolution, edge detection, shape analysis, texture analysis, and recognition. This includes pyramids, wavelets, 2D shape description through contour primitives, and deformable templates (e.g., 'snakes'). Video processing concentrates on motion analysis. This includes the motion estimation methods, e.g., optical flow and block-based methods, and motion segmentation. The course emphasizes experimenting with the application of algorithms to real images and video. Students are encouraged to apply the algorithms presented to problems in a variety of application areas, e.g., synthetic aperture radar images, medical images, entertainment video image, and video compression.
Section P is for Portugal students only.