Research Information System
IEEE-EURASIP Workshop on Nonlinear Signal and Image Prcessing (NSIP 99), Antalya, Turkey, 20 - 23 June 1999, pp.321-325
This paper presents a multiple classifier system for text-independent speaker identification (SI). For the speaker identification problem, several different classifiers can be developed, each having strengths and weaknesses compared to the others. When the strengths and weaknesses of the individual classifiers do not overlap, i.e. a speaker which is misclassified by one classifier is correctly classified by some others, robust classification systems can be developed with the use of multiple classifiers. The studies in multiple classifier systems mainly concentrate on reliable methods of extracting, contextual information (i.e. strengths and weaknesses) about the classifiers and the methods of combining these classifiers. In this paper, a method is proposed for the extraction of contextual information about the classifiers and a rule based approach is developed for the combination of the information from different classifiers.