Digital Signal Processing
Key researchers: Professor Costas Xydeas
Research is focused on the development of novel signal processing algorithms and systems and their efficient operation in different application areas. Research focuses on speech and image processing, on-line signal modelling and predicting behaviours, classification and prediction, evolving adaptive fuzzy systems, information fusion, data mining and knowledge extraction, generation of photorealistic quality textures on 3D surface models and multimedia Information Systems.
InfoLab21 researchers are internationally renowned for their speech analysis, synthesis and compression work and have developed the first UK prototype of a general multimedia database and also demonstrated the applicability of content retrieval of raster images using automatic interpretation of images. Pioneering research into evolving fuzzy rule-based modelling techniques forms the basis for designing next generation, embedded systems with evolving intelligence. Advances made in speech coding, noise cancellation and fusion technologies have resulted in systems and solutions adopted by companies such as AMD, Anite, ULTRA, and BAE.
Case Study
Amplifiable Bi-directional Texture Functions for 3D High Fidelity Images
This project is an investigation into advanced techniques that use textures to enhance the exhibition of surface details in 3D computer generated images and aims to bridge the existing gap observed in the equality of real world images and computer synthesized images. The project has a wide range of potential applications, e.g. computer games, computer animation, broadcast television, mobile communication, Web design, etc., all of which are facing demands for improved image quality and thus take full advantage of continual advances in graphics/visualisation and display technologies.
Case Study
Collaborative autonomous agents with evolvable intelligence
This program, in collaboration with Ford Motor Co. and supported by J&S Marine (an underwater communications business, and Associate Company) aims to develop and implement autonomous agents offering advanced functionalities of classification, modelling, prediction and decision making. These systems are incorporated into ActiveMedia robot systems that perform tasks in previously unseen static or dynamically changing environments, first as single entities and then in a team.
