Matched filter in the low-number count Poisson noise regime: an efficient and effective implementation

TL;DR

This paper introduces an efficient saddle point approximation method for computing the false alarm probability of the matched filter in low-count Poisson noise, improving over previous simulation-based approaches.

Contribution

It provides a novel analytical approach using saddle point approximation for the matched filter in low-count Poisson noise, enhancing computational efficiency.

Findings

The saddle point approximation accurately estimates false alarm probabilities.

The method reduces computational time compared to simulation-based approaches.

Limitations of the matched filter in practical low-count Poisson scenarios are discussed.

Abstract

The matched filter (MF) is widely used to detect signals hidden within the noise. If the noise is Gaussian, its performances are well-known and describable in an elegant analytical form. The treatment of non-Gaussian noises is often cumbersome as in most cases there is no analytical framework. This is true also for Poisson noise which, especially in the low-number count regime, presents the additional difficulty to be discrete. For this reason in the past methods have been proposed based on heuristic or semi-heuristic arguments. Recently, an analytical form of the MF has been introduced but the computation of the probability of false detection or false alarm (PFA) is based on numerical simulations. To overcome this inefficient and time consuming approach we propose here an effective method to compute the PFA based on the saddle point approximation (SA). We provide the theoretical framework and support our findings by means of numerical simulations. We discuss also the limitations of the MF in practical applications.