stem/AI/Neural Networks/CNN/FCN/FCN.md
andy dcc57e2c85 vault backup: 2023-05-31 17:33:05
Affected files:
.obsidian/graph.json
.obsidian/workspace-mobile.json
.obsidian/workspace.json
Gaming/Steam controllers.md
History/Britain.md
STEM/AI/Neural Networks/CNN/CNN.md
STEM/AI/Neural Networks/CNN/FCN/FCN.md
STEM/AI/Neural Networks/CNN/FCN/ResNet.md
STEM/AI/Neural Networks/CV/Datasets.md
STEM/AI/Neural Networks/Properties+Capabilities.md
STEM/AI/Neural Networks/Transformers/Attention.md
STEM/AI/Properties.md
Tattoo/Engineering.md
Tattoo/Sources.md
Tattoo/img/snake-coil.png
Untitled.canvas
2023-05-31 17:33:05 +01:00

42 lines
943 B
Markdown

Fully [[Convolution]]al Network
[[Convolutional Layer|Convolutional]] and [[UpConv|up-convolutional layers]] with [[Activation Functions#ReLu|ReLu]] but no others (pooling)
- All some sort of Encoder-Decoder
Contractive → [[UpConv]]
# Image Segmentation
- For visual output
- Previously image $\rightarrow$ vector
- Additional layers to up-sample representation to an image
- Up-[[convolution]]al
- De-[[convolution]]al
![[fcn-uses.png]]
![[fcn-arch.png]]
# Training
- Rarely from scratch
- Pre-trained weights
- Replace final layers
- [[MLP|FC]] layers
- White-noise initialised
- Add [[upconv]] layer(s)
- Fine-tune train
- Freeze others
- Annotated GT images
- Can use summed per-pixel log [[Deep Learning#Loss Function|loss]]
# Evaluation
![[fcn-eval.png]]
- SDS
- Classical method
- 52% mAP
- FCN
- 62% mAP
- Intersection over Union
- IOU
- Jaccard
- Averaged over all images
- $J(A,B)=\frac{|A\cap B|}{|A\cup B|}$