Abstract: Transfer learning is a proven technique in 2D computer vision to leverage the large amount of data available and achieve high performance with datasets limited in size due to the cost of ...
Abstract: Supervised methods on 3D medical image segmentation need large amounts of annotated data, but annotating is time-consuming. Also, existing 3D segmentation methods capture more global ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback