Multivariate discrimination in quantum target detection

Abstract

We describe a simple multivariate technique of likelihood ratios for improved discrimination of signal and background in multi-dimensional quantum target detection. The technique combines two independent variables, time difference and summed energy, of a photon pair from the spontaneous parametric downconversion source into an optimal discriminant. The discriminant performance was studied using experimental data and Monte Carlo modeling with clear improvement shown compared to previous techniques. As novel detectors become available, we expect this type of multivariate analysis to become increasingly important in multi-dimensional quantum optics.