Transverse phase space tomography in an accelerator test facility using image compression and machine learning

We describe a novel technique, based on image compression and machine learning, for transverse phase space tomography in two degrees of freedom an accelerator beamline. The technique has been used the CLARA test facility at Daresbury Laboratory: results from learning method are compared with those conventional algorithm (algebraic reconstruction) applied to same data. use allows reconstruction 4D distribution beam be carried out much more rapidly than using algorithms also enables reduce significantly size data sets involved analysis. Results least as good algebraic characterizing behavior, terms variation response quadrupole strengths.