Statements (60)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:television_channel
|
| gptkbp:application |
Medical Image Segmentation
|
| gptkbp:architectural_style |
gptkb:Alumni_Association
|
| gptkbp:benefits |
Robustness to noise
Better localization Improved segmentation accuracy |
| gptkbp:contribution |
Reduced overfitting
Enhanced gradient flow Improved feature representation |
| gptkbp:developed_by |
Zhou et al.
|
| gptkbp:features |
Attention mechanisms
Residual connections Adaptive feature fusion Deep supervision Hierarchical feature learning Multi-scale feature extraction Nested architecture Nested skip pathways Skip pathways at multiple levels |
| gptkbp:field_of_study |
gptkb:viewpoint
gptkb:Deep_Learning Medical Imaging |
| gptkbp:game_components |
Decoder
Encoder Skip connections |
| https://www.w3.org/2000/01/rdf-schema#label |
Nested U-Net
|
| gptkbp:impact |
Widely cited in literature
High impact in medical imaging research |
| gptkbp:improves |
gptkb:Standard_U-Net
|
| gptkbp:input_output |
Medical images
Segmentation maps Variable input size |
| gptkbp:is_a_tool_for |
gptkb:Graphics_Processing_Unit
gptkb:Keras gptkb:Py_Torch |
| gptkbp:is_compared_to |
gptkb:Attention_U-Net
gptkb:Seg_Net FCN |
| gptkbp:is_evaluated_by |
Specificity
Sensitivity Pixel accuracy |
| gptkbp:losses |
Cross-entropy loss
Dice loss |
| gptkbp:performance |
Jaccard Index
Dice Coefficient |
| gptkbp:provides_information_on |
gptkb:ISIC_2017
|
| gptkbp:published_year |
gptkb:2018
|
| gptkbp:resolution |
Variable output size
|
| gptkbp:training |
Data augmentation
Transfer learning Fine-tuning End-to-end training Large annotated datasets |
| gptkbp:use_case |
Cell segmentation
Lesion segmentation Organ segmentation Tumor segmentation Vascular segmentation |
| gptkbp:bfsParent |
gptkb:Alumni_Association
|
| gptkbp:bfsLayer |
3
|