Intelligence in Modern Communication Systems

Modern communication systems encompass noise suppression algorithms to compensate for recording and transmission errors as well as multimedia data representation modules offering both expressive and discriminating capabilities. Fusion of decisions derived by mono-modal detectors and features extracted separately for each data type should be supported. In addition, modern terminals should be equipped with humancentered interfaces and should authenticate the user on the basis of biometric information. Moreover, the design of copyright protection procedures is needed. In the paper, state-of-the-art techniques are reviewed and applications are described. The potentialities of neural networks to provide the desired intelligence are highlighted as well.