mett29/DL-Competition
Repo for the 'Artificial Neural Networks and Deep Learning' competition - 2019/2020
Jupyter Notebook
Issues
- 1
Overlay di patch
#24 opened by DanieleParravicini - 1
Set building threshold less than 0.5
#25 opened by DanieleParravicini - 0
Change dice loss weight
#26 opened by DanieleParravicini - 2
- 2
Randomly shifting masks
#30 opened by DanieleParravicini - 1
Consider adding/removing buildings
#31 opened by DanieleParravicini - 1
Image preprocessing
#14 opened by mett29 - 5
Weights on area
#20 opened by DanieleParravicini - 3
Weight based on distance
#21 opened by DanieleParravicini - 0
Loss function
#32 opened by DanieleParravicini - 0
IOU error
#29 opened by DanieleParravicini - 0
- 4
Close /open on results
#18 opened by DanieleParravicini - 0
training/evaluation on Patch
#23 opened by DanieleParravicini - 0
Prior on shape
#22 opened by DanieleParravicini - 0
Transfer learning
#19 opened by DanieleParravicini - 0
Is SegNet better?
#17 opened by mett29 - 0
Dropout and Conv2DTranspose
#16 opened by mett29 - 0
Borders as a new layer of information
#15 opened by mett29 - 1
Implement crossvalidation
#4 opened by mett29 - 2
Prepare python notebook
#11 opened by DanieleParravicini - 0
Make a function that produces a json (document) training split that has to be sent to instructors
#13 opened by DanieleParravicini - 2
Model with skip connection
#9 opened by DanieleParravicini - 1
Consider not resizing the images
#2 opened by mett29 - 1
Add gaussian noise
#10 opened by DanieleParravicini - 1
Fit weight
#8 opened by DanieleParravicini - 0
Commit!!
#12 opened by DanieleParravicini - 1
Weigth decay
#6 opened by DanieleParravicini - 2
Join original and augmented datasets
#1 opened by mett29 - 0
Adding callbacks
#3 opened by mett29 - 0
Weight dataset sampling
#7 opened by DanieleParravicini - 0
Confusion Matrix
#5 opened by mett29