issues
search
Jbartlett6
/
SDNet
SDNet, a model based deep learning method to estimate high quality FODs from indersampled DWI signal data.
2
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Results should be stored correctly every time inference/evaluation is run
#15
Jbartlett6
opened
1 year ago
2
Cannot add MAE error correctly to only take into consideration voxels where there are fixels - would normally use -ignorezeros but the values we want to ignore are actually pi
#14
Jbartlett6
closed
1 year ago
1
Perform Inference flag in the eval.py script doesn't work
#13
Jbartlett6
closed
1 year ago
0
eval requires training details even when perform inference is False, in this case it shouldn't be neccessary since there is no need to load the network
#12
Jbartlett6
closed
1 year ago
0
Add some kind of reminder/manual catch for when continuing training to prevent unwanted overwriting occurring.
#11
Jbartlett6
closed
1 year ago
0
Add an inference to the options.py script.
#10
Jbartlett6
closed
1 year ago
0
Add the MAE to the eval function.
#9
Jbartlett6
closed
1 year ago
0
Create a new dataset which takes the undersampled FODs as input for networks rather than the signals.
#8
Jbartlett6
closed
1 year ago
1
Extract performance measures from MRStats output once after ACC or FBA has been performed.
#7
Jbartlett6
closed
1 year ago
1
Adjust the FOD/CSD net combination to take the full under sampled FOD as input.
#6
Jbartlett6
closed
1 year ago
0
Adding FOD 'signal' loss option
#5
Jbartlett6
closed
2 years ago
0
Write a python script to do all of the pre-processing to the dataset once it has been downloaded from aws.
#4
Jbartlett6
closed
1 year ago
0
Refine the tensorboard plots and data information storage specifically for continuing training.
#3
Jbartlett6
closed
1 year ago
3
Complete the argument parser.
#2
Jbartlett6
closed
1 year ago
0
Subject 101410 patch issue - mask is too close to the edge of the image.
#1
Jbartlett6
closed
2 years ago
3