X-Sifter: detecting transients in X-ray data using the optimal Poisson matched filter
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.
Keywords
Cite
@article{arxiv.2412.07858,
title = {X-Sifter: detecting transients in X-ray data using the optimal Poisson matched filter},
author = {Maayane T. Soumagnac and Eran O. Ofek and Shachar S. Israeli and Guy Nir and Imri A. Dickstein},
journal= {arXiv preprint arXiv:2412.07858},
year = {2024}
}
Comments
12 pages, 9 Figures. Submitted to PASP