FCN (Fully Convolutional Networks)
GPTKB entity
Statements (59)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:neural_networks
|
| gptkbp:applies_to |
autonomous driving
medical image analysis satellite image processing |
| gptkbp:architecture |
convolutional layers
deconvolutional layers |
| gptkbp:can_be_combined_with |
CRFs
|
| gptkbp:can_be_extended_by |
3 D segmentation
|
| gptkbp:can_handle |
variable input sizes
|
| gptkbp:developed_by |
Long et al.
|
| gptkbp:has_achieved |
state-of-the-art results
|
| https://www.w3.org/2000/01/rdf-schema#label |
FCN (Fully Convolutional Networks)
|
| gptkbp:improves |
image segmentation tasks
|
| gptkbp:input_output |
pixel-wise predictions
segmentation maps |
| gptkbp:inspired_by |
gptkb:Res_Net
gptkb:VGGNet gptkb:Alex_Net |
| gptkbp:introduced_in |
gptkb:2015
|
| gptkbp:is_based_on |
CNN architecture
|
| gptkbp:is_compared_to |
traditional CNNs
|
| gptkbp:is_evaluated_by |
pixel accuracy
Io U metric mean accuracy |
| gptkbp:is_implemented_in |
gptkb:Tensor_Flow
gptkb:Py_Torch |
| gptkbp:is_optimized_for |
stochastic gradient descent
|
| gptkbp:is_popular_in |
computer vision community
|
| gptkbp:is_related_to |
gptkb:Na'vi
gptkb:Seg_Net |
| gptkbp:is_trained_in |
backpropagation
|
| gptkbp:is_used_in |
gptkb:sports_team
augmented reality environmental monitoring smart cities gesture recognition image classification object detection video analysis facial recognition traffic monitoring scene understanding image restoration video surveillance land cover classification optical character recognition object tracking instance segmentation robot vision style transfer face segmentation |
| gptkbp:replaces |
fully connected layers
|
| gptkbp:requires |
large datasets
|
| gptkbp:used_for |
semantic segmentation
|
| gptkbp:utilizes |
skip connections
|
| gptkbp:bfsParent |
gptkb:Na'vi
gptkb:Deep_Lab gptkb:Seg_Net |
| gptkbp:bfsLayer |
5
|