Statements (91)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:software_framework
|
| gptkbp:application |
gptkb:Photographer
autonomous driving satellite image analysis |
| gptkbp:applies_to |
gptkb:Photographer
gptkb:robot satellite imagery autonomous driving |
| gptkbp:architectural_style |
encoder-decoder architecture
encoder-decoder |
| gptkbp:based_on |
VG G16
|
| gptkbp:developed_by |
gptkb:University_of_Cambridge
|
| gptkbp:enhances |
feature extraction
|
| gptkbp:features |
pixel-wise classification
|
| gptkbp:first_introduced |
gptkb:2015
|
| gptkbp:has |
pre-trained models
multiple variants low memory footprint skip connections up-sampling layers |
| gptkbp:has_achievements |
real-time performance
|
| gptkbp:has_programs |
gptkb:musician
augmented reality image editing object detection |
| gptkbp:has_variants |
Seg Net-B, Seg Net-S
|
| https://www.w3.org/2000/01/rdf-schema#label |
Seg Net
|
| gptkbp:improves |
semantic segmentation
|
| gptkbp:input_output |
images
RGB images segmented images |
| gptkbp:is |
open-source
scalable widely adopted a benchmark model used in academic research used in industry applications effective for urban scene understanding part of the computer vision toolkit suitable for large-scale datasets |
| gptkbp:is_compared_to |
gptkb:Alumni_Association
gptkb:FCN_(Fully_Convolutional_Networks) gptkb:Deep_Lab Fully Convolutional Networks |
| gptkbp:is_evaluated_by |
gptkb:Cam_Vid_dataset
gptkb:Cityscapes_dataset gptkb:Pascal_VOC_dataset mean Intersection over Union (m Io U) |
| gptkbp:is_implemented_in |
gptkb:Graphics_Processing_Unit
gptkb:Keras gptkb:Py_Torch |
| gptkbp:is_open_source |
gptkb:battle
|
| gptkbp:is_optimized_for |
gptkb:military_unit
accuracy memory efficiency |
| gptkbp:is_part_of |
computer vision algorithms
|
| gptkbp:is_popular_in |
computer vision community
|
| gptkbp:is_related_to |
image classification
semantic segmentation deep neural networks instance segmentation |
| gptkbp:is_used_for |
image segmentation
|
| gptkbp:is_used_in |
gptkb:robot
augmented reality applications real-time video processing |
| gptkbp:losses |
softmax loss
|
| gptkbp:performance |
real-time applications
resource-constrained environments segmentation tasks mean Intersection over Union (m Io U) |
| gptkbp:provides |
multi-class segmentation
|
| gptkbp:provides_information_on |
labeled datasets
|
| gptkbp:related_to |
semantic segmentation
convolutional neural networks (CN Ns) |
| gptkbp:requires |
large datasets
GPU for training |
| gptkbp:resolution |
same as input size
|
| gptkbp:suitable_for |
resource-constrained environments
|
| gptkbp:supports |
transfer learning
multi-class segmentation |
| gptkbp:training |
gptkb:CEO
gptkb:Cam_Vid_dataset gptkb:Cityscapes_dataset gptkb:Pascal_VOC_dataset |
| gptkbp:uses |
batch normalization
dropout layers softmax activation function |
| gptkbp:utilizes |
max pooling indices
|
| gptkbp:bfsParent |
gptkb:television_channel
gptkb:Alumni_Association |
| gptkbp:bfsLayer |
3
|