Enrichment of semantic metadata based on interactions with user associations

Abstract

More and more aitdio-visual injormatiun is available on the World Wide Web. There is an increasing amount of broadcast channeb available. In such a situation. to attoin a competitive advantage, contenl ~ e p o ~ i t < ~ t k ~ must progress to collecting and managing knowledge abort! all @pes of injbrmation including films and TV programs. Semantic metadata is the most valuable one for the user who want.s to get knowledge. I t is usuall,v created by a professional analyst in the labor-intensive activiy. A new approach to the creation of semantic metadata inchiding affective one is sugge.sted. It is ba.sed on the collection of use‘s ’ evahtation of aiidiovisual content. A cooperative interaction of a content repositop and users associations resnlts in the cooperative metadata enrichment. The main features of this method are: continuous evaluation in the course of time made hy users msociations with varying background. education and motivation: evaluation of different attributes of audiovisual events on dilf.;rent scales. Experiments were carried out to examine the approach sirgge.sted in the ciirrent paper. The aim of these experiments was to evaluate the emotional tension created bv ,fiattire films ’ events and to select keyfi-ames ofthese events. The statistically proven reszilts .f‘ these experiments confirmed that itset..s with d;jfkrent age f rom children to adults) and vanCng background, education and mativattoti can indicate emotional tension on different scales (scales based on coniple.rit?, or informational attvilurtes, on the degrees of interest and un1iredictabilih, of dramatic ewent.~). The s t o p shapes made on the basis of these e.yperimenta1 results look like ones made l ~ y profi.s.siona1 analysts. However regular dependences between users ’ pmJ;Ie.s and their prefi.rence.s were detected. It means that this method can he used to gficientlv create dif/^erent types of summaries for dljfkrent ypes of TV viewers and the Internet use.’. Index Terms Audiovisual content, metadata enrichment, semantic metadata.