The catalogue was created based on the extended candidate catalogue of the Planck clusters (SZcat) and deep learning algorithm, that was trained on the ACT+Planck maps (Naess et al. 2020).
The ComPACT catalogue contains 2,962 candidates. Below we describe columns:
- Name: ID of a ComPACT candidate
- RA: Right Ascension in decimal degrees (J2000) of maximum pixel
- DEC: Declination in decimal degrees (J2000) of maximum pixel
- S: Object mask area in pixels
- pmax: Maximum probability for an object
- SZcat: Name of the object from the SZcat catalogue
- ACT: Cluster name in the ACT DR5 catalogue
- PSZ2: PSZ2 source name
- Priority: reliability of candidate along S area:
- 1: S > 30 (
$Purity_{min} = 0.84$ ) - 2: S > 25 (
$Purity_{min} = 0.78$ ) - 3: S > 20 (
$Purity_{min} = 0.74$ )
- 1: S > 30 (
For columns we used catalogues:
- SZcatgen: data, Meshcheryakov et al. 2022
- ACT DR5: data, Hilton et al. 2021
- PSZ2: data, Planck Collobaration
Description: Cluster calalogue: ComPACT.csv (v2.0)
- v2.0 Add 'Priority' column, which is responsible for subsamples with different purity and completeness characteristics. Also, We keep the nearest object in 5 arcmin window (before all objects in 5 arcmin window). Also, now we cross-match objects from full catalogue with SZcat, before we crop 5 arcmin window from probability map and analyse groups
- v1.1 Negative RA coordinates in catalog are fixed (e.g -152.41666 -> 207.58333)
- v1.0 Initial release (in folder v1.0)