Identifying Astrophysical Anomalies in 99.6 Million Source Cutouts from the Hubble Legacy Archive Using AnomalyMatch

This repository contains the data released in the paper "Identifying Anomalies in 99.6 Million Source Cutouts from the Hubble Legacy Archive Using AnomalyMatch" (DOI: To be Added on Publication ; ArXiv: 2505.03508).

We release the images and descriptions of every anomalous object found in this paper (1,176 unique objects across 18 different anomaly classifications). The main result of the paper, the catalogue of anomalous objects, is contained in DOR_PG+2025_anomalies.csv. This catalogue contains the SourceID, Right Ascension, Declination, classification and the "Status" of each object. The status shows whether an object has appeared in the astrophysical literature prior to this work. This main catalogue describes every object we found.

We also include the images that correspond to each catalogue. These are contained within the .zip files in the repository. The individual filenames are the SourceIDs of the corresponding objects in the main CSV file.

Please note: These images have been saved as 150x150 gray images using only the F814W filter of the HST. Different dimension images/multi-band photometry will have to be found by the user.

Please also note: Each sub-category of object has only been classified visually. Therefore, for unreferenced objects, these should be considered candidate objects.

A list of the anomaly classifications we have found are below:

1. agn - Zip of sources containing an Active Galactic Nuclei (8 sources).
2. arc - Zip of sources containing a gravitational arc (cluster lensing; 39 sources).
3. candidate strong lens - Zip of sources containing a gravitational lense (galaxy - galaxy lensing; 140 sources).
4. clumpy - Zip containing images of a clumpy galaxy (11 sources).
5. collisional ring - Zip containing a collisional ring galaxy (2 sources).
6. edge-on protoplanetary disk - Zip of sources which contain edge on planetary disks (2 sources).
7. globular cluster - Zip containing sources showing globular cluster (1 source).
8. high-z - Zip of sources containing high redshift galaxies (28 sources).
9. jellyfish - Zip of sources containing a jellyfish galaxy (37 sources).
10. jet - Zip of sources containing a galaxy with a jet (13 sources).
11. lensed quasar - Zip containing images of a (quadruply-) lensed quasar (5 sources).
12. merger - Zip of sources containing an interacting or merging galaxy (629 sources)
13. odd - Zip containing sources with odd morphology (43 sources).
14. overlap - Zip of sources which contain two galaxies which overlap by projection, but are not interacting (39 sources).
15. ring - Zip of sources which contain ring galaxies (12 sources).
16. submillimetre - Zip of sources which contain submillimetre or dusty galaxies (1 source).
17. supernova - Zip of sources which contain a supernova remnant (1 source).
18. unknown - Zip of sources which could not be classified visually (43 sources).

Version History:

1. Version 1: Pre-release of paper and catalogue on ArXiv.
2. Version 1.5: Catalogue and paper submitted to A&A (Changes: revised strong lensing candidate catalogue based on feedback from community).
3. Version 1.6: Updated catalogue and paper after referee comments (Changes: updated paper, and included unknown.zip file in release).