Learning a temporally invariant representation for visual tracking | IEEE Conference Publication | IEEE Xplore