DM-HW2

The homework aims to classify different types of intracerebral hemorrhage with the CT images.

Hardware

The following specs were used to create the original solution.

Ubuntu 18.04 LTS
Intel(R) Core(TM) i7-6700 CPU @ 3.40 GHz
NVIDIA GeForce GTX TITAN X

Installation

All requirements should be detailed in requirements.txt. Using Anaconda is strongly recommended. {envs_name} is the new environment name which you should assign.

conda create -n {envs_name} python=3.7
source activate {envs_name}
pip install -r requirements.txt

Dataset Preparation

You can download the data on the following google drive:

Training data: https://drive.google.com/file/d/1xd7gpJjJ9rJy8XqW1ArfAtkzXr1rvroL/view?usp=sharing

Testing data: https://drive.google.com/file/d/1xd7gpJjJ9rJy8XqW1ArfAtkzXr1rvroL/view?usp=sharing

Prepare Images

After downloading, the data directory is structured as:

TrainingData
  +- epidural
    +- ID_0a5b19112.jpg
    +- ID_0a21c7cde.jpg
    ...
  +- healthy
    +- ID_0a0f3abd0.jpg
    +- ID_0acc9d2bf.jpg
    ...
  +- intraparenchymal
    +- ID_00a1d04a4.jpg
    +- ID_0a1dc9169.jpg
    ...
  +- intraventricular
    +- ID_0a5db43bf.jpg
    +- ID_0a729be82.jpg
    ...
  +- subarachnoid
    +- ID_0a0b55bbd.jpg
    +- ID_0a7ba802a.jpg
    ...
  +- subdural
    +- ID_0a4a21efb.jpg
    +- ID_0a16f9f35.jpg
    ...
+- TestingData
  +- Test_001.dcm
  +- Test_002.dcm
  ...

Data Preprocessing

First, it will transfer the dicom file to jpg file. And, it is going to do the data augmentation. Finally, it is going to split the data randomly to generate a training data and valid data in the input directory. The ratio of the training data and valid data is 8 : 2

$ python3 preprocessing.py

Training

You can do training by following

$ python3 training.py

Make Submission / Testing

You can do testing by following

$ python3 testing.py

nelson870708/DM-HW2