… proportions for fine-tuning MSRAFT, as its training is less time-consuming than MSRAFT+’s
training. … As training supervision we used the multi-scale multi-iteration loss of MS-RAFT. As …

… To this end, we introduce MS-RAFT+: a novel recurrent multi-scale architecture based on
RAFT that … It employs a coarse-to-fine estimation to enable the use of finer resolutions by useful …

… However, one of the currently most successful approaches – RAFT – … that multi-scale
ideas are still valuable. More precisely, using RAFT as a baseline, we propose a novel multi-scale …

… Although multi-scale concepts have recently proven useful for recurrent network … Hence,
based on a single-scale recurrent scene flow backbone, we develop a multi-scale approach that …

… We then propose a raft cultivation area (RCA) extraction method for high spatial resolution
(HSR) remote sensing imagery based on multi-scale feature fusion and spatial rule induction. …

… neural network is to extract high latitude features and does not bring in multi-scale features,
for this reason, we designed MSBlock to accomplish the fusion of multi-scale features. …

… We introduce a meticulously designed multi-scale feature correction block within our MSF-RAFT
framework, aimed at effectively managing and integrating multiple feature streams. As …

… and small objects need a high-resolution network to be … multi-scale network more easily
visualize its decisions from different scale features. In summary, our DS-HCN has high precision …

… (CRAFT) for gridded crop modeling and yield forecasting along with risk analysis and climate
impact studies. CRAFT is … CRAFT requires gridded input data to run the crop simulations on …

… highresolution multi-scale networks. In this paper, we hence propose CCMR: a high-resolution …
, 40]), our multiscale baselines MS-RAFT and MS-RAFT+ do not consider this additional …