Learning visual models of semantic concepts | IEEE Conference Publication | IEEE Xplore