jiaxuan-li / mgu-net Goto Github PK
View Code? Open in Web Editor NEWMulti-Scale GCN-Assisted Two-Stage Network for Joint Segmentation of Retinal Layers and Disc in Peripapillary OCT Images
License: MIT License
Multi-Scale GCN-Assisted Two-Stage Network for Joint Segmentation of Retinal Layers and Disc in Peripapillary OCT Images
License: MIT License
First of all, thanks Jiaxuan-Li for your work and how tidy the code base is. It worked for me almost straight off the shelf!
Regarding the public dataset - it is not in the format the code expects, but rather in Matlab objects. I guess you've been using a script to transform it. Is that the case? If so, would it be possible to share it?
Thanks!
Dan
In the Dice loss, there's a line:
input = torch.exp(input)
but shouldn't it be the sigmoid function? The input is the last conv layer output from the network and a sigmoid will convert it to probability within [0,1]
input = torch.sigmoid(input)
(MGU-Net-main) E:_BACKUP\lbc\MGU-Net-main>python main_ts.py --name tsmgunet -d ./data/dataset --batch-size 1 --epoch 50 --lr 0.001
torch version: 1.10.0
Total amount of train images is : 96
Total amount of eval images is : 96
Total amount of test images is : 96
data_dir : ./data/dataset
name : tsmgunet
workers : 2
step : 20
batch_size : 1
epochs : 50
lr : 0.001
lr_mode : step
momentum : 0.9
weight_decay : 0.0001
t : t1
model_path :
############### tsmgunet ###############
[2021-12-03 04:27:43,359 main_ts.py:278 train_seg] Epoch: [0]
Traceback (most recent call last):
File "main_ts.py", line 394, in
main()
File "main_ts.py", line 386, in main
train_seg(args,train_result_path,train_loader,eval_loader)
File "main_ts.py", line 280, in train_seg
loss,dice_train,dice_1,dice_2,dice_3,dice_4,dice_5,dice_6,dice_7,dice_8,dice_9,dice_10 = train(args,train_loader, model,criterion1, criterion2, optimizer,epoch)
File "main_ts.py", line 53, in train
output_seg1,_,output_seg = model(input_var1)
File "E:\anaconda\envs\MGU-Net-main\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "E:\anaconda\envs\MGU-Net-main\lib\site-packages\torch\nn\parallel\data_parallel.py", line 166, in forward
return self.module(*inputs[0], **kwargs[0])
File "E:\anaconda\envs\MGU-Net-main\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "E:_BACKUP\lbc\MGU-Net-main\models\nets\TSNet.py", line 16, in forward
out1 = self.stage1(inputs)
File "E:\anaconda\envs\MGU-Net-main\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "E:_BACKUP\lbc\MGU-Net-main\models\nets\MGUNet.py", line 106, in forward
up3 = self.up_concat3(center, conv3)
File "E:\anaconda\envs\MGU-Net-main\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "E:_BACKUP\lbc\MGU-Net-main\models\utils\utils.py", line 91, in forward
outputs0 = torch.cat([outputs0,input[i]], 1)
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 52 but got size 53 for tensor number 1 in the list.
I would like to ask what went wrong
test_dataset的参数phase是不是应该设置为predict更加的合理,因为设置test或eval实际上调用的是一个函数
First of all, thank you for your contribution, Jiaxuan Li. Since I have not been able to get a response after filling out the data collection form, I would like to ask if you can give me a way to get the dataset. My Google email is [email protected].
Thanks!
Tom
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.