OSICpypy

Python project as a proposal for the Kaggle competition OSIC Pulmonary Fibrosis Progression available at https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression

Core ideas to get the proposed solution

1. Image processing:
   • Dicom images are processed using the python library known as imageio.
   • An index has been designed based on the training data; see the file Project/Support-Sourced/generic_imageio.py
2. Predicting: Prediction is based on a stacking solution fostered by scikit-learn known as Combine predictors using stacking, that scikit-learn solution has been adequated to predict; see the file Project/MLModel/MachineLearningModel_PydicomFeatures.py

Structure of the project

see the use case diagram powered by Lucidchart

Activity diagram

see the activity diagram powered by Lucidchart

Conditionning and Installation guide

1. Conditionning | Phase 1: Get the repository by commiting git clone https://github.com/mazalgarab-git/OSICpypy.git
2. Conditionning | Phase 2: Get libraries by commiting pip install -r requirements.txt
3. Conditionning | Phase 3: Build a root Y:/Kaggle_OSIC/
4. Conditionning | Phase 4: Get basic structure (i.e. tree-directory structure to start based on the file Data/TreeStructure_To_start.txt) by unzipping Data/TreeStructure_To_start.rar into Y:/
5. Conditionning | Phase 5: Get Kaggle data (i.e. pydicom files on test and train directories) from https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/data
6. Conditionning | Phase 6: Put Pydicom data as follows: (1) train dataset to Y:/Kaggle_OSIC//2-Data/train/; (2) test dataset to Y:/Kaggle_OSIC/2-Data/test/
7. Installation | Phase 1: Put files on Project/ into Y:/Kaggle_OSIC/OSICpypy/

Remark

1. The submissions available at OSICpypy/submissions_to_2020_11_03/ were not submitted to compete.