Meade AD
ORCID: 0000-0003-0353-9768
· 1 paper in corpus
Identifying image features that associate strongly with diagnostic or prognostic classes in large-scale, multi-channel spatial imaging is challenging without pixel-level annotations. We present DREAMER-S, an attention-based multiple-instanc…