Search papers, labs, and topics across Lattice.
1
0
3
30
Pretrained vision transformers can achieve state-of-the-art medical image segmentation with significantly less labeled data, outperforming fully supervised convolutional networks even when trained on only 25% of the dataset.
