A newmassfunctions assignment intheDempster-Shafer theory: thefuzzy statistical approach

This paper describes anewanalysis oftheevidential rea- soning basedonthefuzzy statistical approach. Theevidential rea- soning usestheDempster-Shafer (DS)theory with anewapproach to assign themassfunctions. Wetest theproposed method onabasic simulated noisy fuzzy imagecomposed onthree pureclasses, andwe compare themassfunctions representation withstandard statistical methods andfuzzy methods. Forareal study, weapply thenewdata fusion process toremote sensing data fusion. astatistical andafuzzy analysis based ontheFSEM(fuzzy stochastic expectation maximization) algorithm. We will test thenewmethod onabasic simulated noisy fuzzy image com- posed onthree pureclasses, andwewill compare themass functions representation foreachcase:thestatistical method withSEM,thefuzzy method withFCM andthefuzzy statisti- calmethod withFSEM.Moreover, wewill apply thenewdata fusion process toremotesensing data fusion.