Abstract:We present X-sifter, a software package designed for near-optimal detection of sources in X-ray images and other forms of photon images in the Poisson-noise regime. The code is based on the Poisson-noise-matched filter (Ofek & Zackay), which provides an efficient method for calculating the delta log-likelihood function for source detection. The software accounts for several complexities inherent in real data, including variations in both the instrumental Point Spread Function (PSF) and background across the detector and as a function of energy. We validate the pipeline using real data with simulated source injections, as well as actual Chandra images. A comparison between the sources detected by our pipeline and those in the Chandra Source Catalog (CSC) suggests an approximate ~30% increase in the number of detected (real) sources. Near the detection limit, the reported S/N of our pipeline is approximately 1.3x higher than that of the CSC. This corresponds to a factor of 1.8 increase in survey speed.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to detect sources more effectively in the low - count regime (such as in X - ray, ultraviolet, γ - ray, and neutrino detectors, etc.). Specifically, the authors have developed a software package named X - Sifter, which aims to perform approximately optimal source detection in X - ray images and other forms of photon images through the optimal Poisson - matched filter (PMF) method. ### Problem Background When processing astronomical images, a source detection step is usually required. For the independent and identically distributed (i.i.d.) Gaussian noise case, an optimal solution already exists, namely the matched filtering method. However, in the low - count regime, photon statistics follow Poisson statistics, and the traditional Gaussian matched filtering method becomes sub - optimal. In this case, the signal - to - noise ratio (S/N) of source detection will be lower than the theoretical maximum. ### Solution To solve this problem, X - Sifter implements the Poisson - matched filter (PMF) formalization method proposed by Ofek & Zackay and has been improved for various complexities in actual data, such as the changes of background and point spread function (PSF) on the detector and with energy. The algorithm improves the detection performance in the following ways: 1. **Optimizing Detection Statistics**: Calculate the log - likelihood difference statistic \( S \) using the Poisson - matched filter and achieve efficient calculation through convolution operations. 2. **Handling Threshold Setting**: Considering the influence of source flux \( F_s \) on the threshold \( S_{th} \), determine appropriate \( F_{th} \) and \( S_{th} \) by numerical methods to ensure optimal performance at the detection limit. 3. **Simulating Background Distribution**: Calculate \( P(S|H_0) \) through numerical simulation to accurately evaluate the false - alarm probability \( \gamma \). 4. **Time - Segmenting Data**: To detect short - timescale transient sources, divide all data into shorter time periods and then apply the search algorithm, thereby improving the detection ability of transient sources. ### Main Contributions - **Improved Detection Performance**: Compared with the Chandra Source Catalog (CSC), X - Sifter can detect approximately 30% more real sources. - **Increased Signal - to - Noise Ratio**: Near the detection limit, the signal - to - noise ratio reported by X - Sifter is about 1.3 times higher than that of CSC, which is equivalent to an 1.8 - fold increase in the survey speed. Through these improvements, X - Sifter provides a more effective source detection method in the low - count regime, significantly enhancing the source detection ability in astronomical research.

X-Sifter: detecting transients in X-ray data using the optimal Poisson matched filter

Optimal matched filter in the low-number count Poisson noise regime and implications for X-ray source detection

Bayesian methods of astronomical source extraction

TransientX: A high performance single pulse search package

TransientX: A high-performance single-pulse search package

eBASCS: Disentangling Overlapping Astronomical Sources II, using Spatial, Spectral, and Temporal Information

Automatic estimation of flux distributions of astrophysical source populations

3D Distance-Weighted Wiener Filter for Poisson Noise Reduction in Sinogram Space for SPECT Imaging

Statistical Algorithms for Identification of Astronomical X-Ray Sources

A fast and portable Re-Implementation of Piskunov and Valenti's Optimal-Extraction Algorithm with improved Cosmic-Ray Removal and Optimal Sky Subtraction

Proper image subtraction - optimal transient detection, photometry and hypothesis testing

HELP: XID+, The Probabilistic De-blender for Herschel SPIRE maps

Virtual Point Source Synthesis method for 3D Scintillation Detector Characterization

A Probabilistic Method of Background Removal for High Energy Astrophysics Data

TRANSLIENT: Detecting Transients Resulting from Point-source Motion or Astrometric Errors

X-CSIT: a toolkit for simulating 2D pixel detectors

SRG /ART-XC Galactic Bulge deep survey. I. Maximum likelihood source detection algorithm for X-ray surveys

TRANSLIENT: Detecting Transients Resulting from Point Source Motion or Astrometric Errors

Image reduction pipeline for the detection of variable sources in highly crowded fields

Principled point-source detection in collections of astronomical images

Efficient Search and Detection of Faint Moving Objects in Image Data