Pattern Analysis

This lecture first supplement the methods of preprocessing presented in Pattern Recognition 1 by some operations useful for image processing. In addition several approaches to image segmentation are shown, like edge detection, recognition of regions and textures and motion computation in image sequences. In the area of speech processing approaches to segmentation of speech signals are discussed as well as vector quantization and the theory of Hidden Markov Models.  Accordingly several methods for object recognition are shown. Above that different control strategies usable for pattern analysis systems are presented and therefore also several control algorithms e.g. the A(star) - algorithm.  Finally some formalisms for knowledge representation in pattern analysis systems and knowledge-based pattern analysis are introduced.