Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Corresponding author
3
0
3
Decoupling lesion detection from text alignment leads to unprecedented accuracy in grounding free-text findings in 3D chest CT scans.
MAGE redefines gastric neoplasm classification by forcing models to prioritize intrinsic morphological features over misleading texture cues, leading to unprecedented accuracy in clinical settings.
Achieving a Dice score of 0.4805 for lesion segmentation reveals the significant challenges posed by heterogeneous lesions in msTBI, while maintaining high specificity for non-lesion identification.