431-469 Multimedia Signal Processing

431-325 Stochastic Signals and Systems

On completion of this subject students should have obtained an appreciation of the rapidly emerging field of pattern recognition and artificial neural networks, understand the application of pattern recognition techniques to problems in signal processing, be able to design intelligent systems for the solution of real-world prediction and classification problems and be able to understand the signal processing techniques needed for the design of multimedia systems and its applications.

Topics include introduction to statistical and neural network based pattern recognition methods; multilayered feedforward neural networks; self organising neural networks; feedback neural networks; applications to signal and image processing; speech, image and video processing; image segmentation and feature extraction; image tracking and recognition systems; and speech and image compression techniques, video on demand, applications of multimedia signal processing.

One 3-hour end-of-semester written examination (70%); mid-semester test and/or project report not exceeding 20 pages including appendices, diagrams, tables, graphs and computer output (30%). The relative weighting of test and/or project report will be specified both in the first lecture and on the subject web page at the start of semester.