S/N maps computed with the PACO algorithm. Left: S/N map obtained with the PACO detection Algorithm ¹. Covariance matrices are regularized by adding the fraction ϵ = 10⁻⁶ of identity matrix (see Eq. (A.1)). Middle: S/N map obtained with the PACO detection Algorithm ¹. Covariance matrices are estimated using the shrinkage principle (see Eqs. (A.3) and (A.8)). Right: S/N map obtained with the PACO estimation Algorithm ³. Covariance matrices are estimated using the shrinkage principle (see Eqs. (A.3) and (A.8)). The first 32 detections are marked on the maps using square patterns. The 60 first detections are plotted as bar charts below each S/N map, ordered by decreasing S/N values, with true detections in pink (true background sources) or blue (injected fake sources), and false detections in red. Circles indicate the location of all faint point sources.