Snapshot of how the various algorithms perform in the evaluation dataset. Top, from left to right: Original, RFit, and DeepFilter AIA 304 Å as well as the IRIS 1400 Å and 2796 Å images produced at a time with quiescent coronal rain. The AIA images use 0.5 power scaling, while the IRIS images use 0.3 power scaling. Middle: Same images binarised by the optimal threshold found in Section 5.3. Bottom: Differences and similarities between the AIA 304 Å and IRIS 1400 Å binarised images (see legend for colour scheme used).