- It is my graduate thesis in university (2019).
- Topic: Detect liver tumors in CT image
- Dataset: Liver tumor Segmentation Challenge (LiTS) contain 131 contrast-enhanced CT images provided by hospital around the world. If you'd like to download it, you can find in this link: https://competitions.codalab.org/competitions/17094
- Used models: CNN3D and UNet3D (3D mean using convolution 3D (3x3x3), max pooling 3D (3x3x3),...)
- I use python==3.7.0 and tensorfow==1.14.0.
- About the library, you can install all packages in requirements.txt by run this command:
pip install -r requirements.txt
-
Open file: Analyse_data.ipynb
-
Function Histogram() to draw the data's histogram. See below chart:
-
Function Information to log the tumor information (size of tumor in whole image)
-
Function PixelRatio to draw the ratio chart between tumor and back ground. See below chart:
- A list of papers refer to this project:
- A list of other great sources that were sources of inspiration: