Concept两阶段框架 / 定位+分类范式 / coarse-to-fine框架
Two-Stage Deep Learning Framework
A hierarchical deep learning paradigm that decomposes medical imaging detection tasks into first localizing suspicious regions and then performing pathological classification of those regions. It is commonly used in medical imaging tasks where lesions are sparse and classes are imbalanced.
Timeline (last 90 days)
Oct 31
The study proposes a two-stage deep learning framework to detect neonatal brain lesions from MRI by grade, with the first stage localizing suspicious regions and the second stage performing pathological classification.
Unverified50%