abhi4ssj / few-shot-segmentation Goto Github PK
View Code? Open in Web Editor NEWPyTorch implementation of 'Squeeze and Excite' Guided Few Shot Segmentation of Volumetric Scans
License: MIT License
PyTorch implementation of 'Squeeze and Excite' Guided Few Shot Segmentation of Volumetric Scans
License: MIT License
Hi,
Thanks for sharing the code with the community. I have one quick question. In convert_h5.py, how do you prepare those corresponding data? i.e., what is the FS mean? Could you leave us with the data preparation instruction? Hopefully, the guidance will make us easier to run your architecture.
Thanks much!
--Ruida
"""
Convert to h5 utility.
Sample command to create new dataset - python utils/convert_h5.py -dd /home/masterthesis/shayan/nas_drive/Data_Neuro/OASISchallenge/FS -ld /home/masterthesis/shayan/nas_drive/Data_Neuro/OASISchallenge -trv datasets/train_volumes.txt -tev datasets/test_volumes.txt -rc Neo -o COR -df datasets/MALC/coronal
- python utils/convert_h5.py -dd /home/masterthesis/shayan/nas_drive/Data_Neuro/IXI/IXI_FS -ld /home/masterthesis/shayan/nas_drive/Data_Neuro/IXI/IXI_FS -ds 98,2 -rc FS -o COR -df datasets/IXI/coronal
- python3.6 utils/convert_h5.py -dd /home/deeplearning/Abhijit/nas_drive/Abhijit/WholeBody/CT_ce/Data/SilverCorpus -ld /home/deeplearning/Abhijit/nas_drive/Abhijit/WholeBody/CT_ce/Data/SilverCorpus -trv datasets/test_volumes_silver.txt -tev datasets/test_volumes_silver.txt -rc WholeBody -o AXI -df datasets/silver_corpus
"""
Hi,
Thanks for sharing the code with the community. I have one quick question. In few_shot_segmentor.py line 81, there should be
e4, _, ind4 = self.encode4(e3)
but not
e4, _, ind4 = self.encode3(e3)
right?
Or is there any special meaning?
Thanks!
Hi,
I am trying to the code on Google Colab but I get this error:
Traceback (most recent call last):
File "/content/drive/MyDrive/sofa/few-shot-segmentation/run.py", line 7, in
import few_shot_segmentor as fs
File "/content/drive/MyDrive/sofa/few-shot-segmentation/few_shot_segmentor.py", line 6, in
from nn_common_modules import modules as sm
File "/usr/local/lib/python3.9/dist-packages/nn_common_modules/modules.py", line 526
else:abdominal_segmentation_2] - 1) / 2)
^
SyntaxError: unmatched ']'
Hi, I'm using my dataset and trying to run this code, but i haven't figured it out, what should the data format look like. It would be very helpful for me if you can share how to adjust the format of the data to run "run.py". Truly thx :P
In paper "Recalibrating Fully Convolutional Networks withSpatial and Channel ‘Squeeze & Excitation’ Blocks", we can find that using scSE block can get better performance ,but this paper didn't do experiments about it, has anyone tried this?
Why does the author only use sSE instead of scSE? Is it because scSE has too many parameters?
Hello! Thank you very much for sharing the code.
Could you share the download link and data processing of the datasets, which will help me a lot. Thank you!
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.