Search papers, labs, and topics across Lattice.
Affiliation:
2
0
3
4
No single foundation model rules 3D radiology: across 2.1 million volumetric scans and 108 benchmarks, ConvNets consistently outperform transformers at dense spatial localization, while transformers decisively win on global semantic reasoning.
A deep learning model can now distinguish between tumor recurrence and radiation effects in brain cancer patients with high accuracy, primarily by leveraging radiotherapy dose distributions alongside MRI data.
