Skip to content

deeplearningrosario/Pediatric_Bone_Age_Challenge

Repository files navigation

Pediatric Bone Age Challenge

This is a first approach to Deep Learning by an investigation group from Rosario, Santa Fe, Argentina. We used another team's solution to submerge ourselves. However, we've since modified it to suit our needs and to apply any improvements we saw fit.

The order of execution would be

  1. pickle_dataset
  2. main
  3. check

Dependencies needed

Install on Linux

sudo pip install keras
sudo apt-get install python3-matplotlib

python3 -mpip install opencv-python --user

python3 -mpip install pandas --user

python3 -mpip install numpy --user

Download DataSet rsna-bone-age

  1. Generate the API key

Go to the Kaggle account, link https://www.kaggle.com/ your user name /account

Click Create New API Token and then save the json file in 'home' (linux users), mode info kaggle-api

  1. Install kaggle cli
pip install kaggle
  1. Move our API key to kaggle path
kaggle
mv ./kaggle.json ~/.kaggle/kaggle.json

or

mv ./kaggle.json /root/.kaggle/kaggle.json
  1. Downloader dataset in our repository
kaggle datasets download -d kmader/rsna-bone-age -p ./
  1. Unzip
unzip boneage-test-dataset.zip

unzip boneage-training-dataset.zip
  1. Move CSV to dataset folder
mv boneage-training-dataset.csv ./boneage-training-dataset

mv boneage-test-dataset.csv ./boneage-test-dataset
  1. Result
.
├── attention_model.py
├── boneage-test-dataset
├── boneage-training-dataset
├── check_no_gender.py
├── check.py
├── dataset_sample
├── .git
├── .gitignore
├── main_no_gender.py
├── main.py
├── pickle_dataset_multiprocessing.py
├── pickle_dataset.py
├── prueba.py
├── README.md
...