Project leader: Ing. Igor Vajda, DrSc.

Department: SI

Supported by (ID): GA102/07/1131

Duration: 1970 - 2009

Abstract:

New theoretically well founded general definitions of the amount of information contained in messages, signals and vision fields from stochastically specified sources. Specification of particular formulas and presentation of properties of the new information measures in a form suitable for engineering applications in research and development of information processing systems. Application of these measures in evaluating upper and/or lower bounds for quality of decisions in the problems of signal detection and identification of parameters of information sources. Development of methods for asymptotically errorless approximation of these measures based on standard statistical data from information sources. Algorithmization of these methods and preparation of a package of computer programs for their practical implementations. Publication of the new concepts, results and approximation procedures in renowned international journals and at the symposia of IEEE.