Top tagging at ATLAS

Studies of the top physics boosted sector have known a fast-growing development with the arrival of high-energy data at LHC. Different techniques to identify high-$p_{\textrm{T}}$ top quarks based on substructure analyses of large radius jets have been developed for Run-1 and Run-2 data. New results are presented on the optimization, performances and performance comparisons, for the different techniques (HEPTopTagger, Shower Deconstruction and substructure variables cut-base taggers), using pp collision data and MC simulations at $\sqrt{s}=8$ TeV in ATLAS