Neural network based cluster reconstruction in the ATLAS pixel detector

The ATLAS Pixel detector is currently determining particle positions at 8TeV proton-proton collisions at the LHC producing jets with high pT. In such dense environments, the separation between particles becomes small, such that the charge deposited by multiple particles is reconstructed as single clsuters. A NN-based clustering algorithm has been developed to identify such merged clusters. By using all cluster information, the NN is ideal to estimate the particle multiplicity and for each of the estimated number of particles, the position with its uncertainty. As a result of the NN reconstruction, the number of hits shared by tracks is strongly reduced. Furthermore, the impact parameter improves by about 15% which indicates boosted prospects for physics analysis.