Electron and photon identification in ATLAS. Comparison between test beam data and simulation

The ATLAS experiment at the Large Hadron Collider (LHC) will face the challenge of efficiently selecting interesting candidate events in $pp$ collisions at 14 TeV center-of-mass energy, whilst rejecting the enormous number of background events. In this talk an overview of the current physics and system performance of the offline selection for electrons and photons is given. Test-beam data, covering a large part of the final detector, have been analysed, and measurement of several important particle identification estimators are presented. Finally, particle identification performance has been evaluated using Monte Carlo simulations. The efficiency for the signal channels as well as the background rejection capability will be highlighted.