Abstract: Data augmentation (DA) stands out as a powerful technique to enhance the generalization capabilities of deep neural networks across diverse tasks. However, in low-level vision tasks, DA ...
Abstract: Medical image datasets often suffer from inadequate labeling, primarily due to ethical constraints, expertise challenges, and high annotation costs. In addition, domain shifts, resulting ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results