Skip to main content

Table 1 DSCs between different weak labels and ground truths of the training dataset

From: Weakly-supervised convolutional neural networks of renal tumor segmentation in abdominal CTA images

 

Bounding boxes

Pseudo masks

Fusion masks

d = 0

0.666

0.862

0.874

d = 5

0.466

0.801

0.810

d = 10

0.341

0.679

0.691

Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us