extract from a time-frequency representation the biggest energy dots
[tfr2,OrderedSurfaces]=tfrsurf(tfr); [tfr2,OrderedSurfaces]=tfrsurf(tfr,threshold); [tfr2,OrderedSurfaces]=tfrsurf(tfr,threshold,keep); [tfr2,OrderedSurfaces]=tfrsurf(tfr,threshold,keep,trace);
time-frequency representation.
the energy threshold, in %
number of dots to keep
if nonzero, the progression of the algorithm is shown (default : 0).
N=256; sig=fmlin(N,0.1,0.3)+fmlin(N,0.3,0.4)+2*fmlin(N,0.05,0.2).*amgauss(N,190,70); tfr=tfrwv(sig,1:N,128); [tfr2,OrderedSurfaces]=tfrsurf(tfr,5,3,1); scf(1);tfrview(tfr,sig,1:N,'tfrwv',[2 1 5 10 128 2 1 0 0 .5]) title('original tfr'); scf(2);tfrview(tfr2,sig,1:N,'tfrwv',[2 1 5 10 128 2 1 0 0 .5]); title('modified tfr'); scf(3);plot(1:10,OrderedSurfaces(1:10),'-',1:10,OrderedSurfaces(1:10),'o'); //semilogy title('number of points of the 10 biggest dots'); See also imextract. | ![]() | ![]() |