Radiomics Extraction and Evaluation based on Binary 3D Segmentation Images

• segmentations include masks of liver structures such tumors, ablations, vessels and other structures
• metrics include the Euclidean Distance between two objects (eg. Ground Truth segmentation of Tumor vs. Predicted Segmentation of Tumor, Segmentation of Ablation vs. Segmentation of Tumor)
• Mauerer et Al. algorithm for calculating euclidean distances between two objects from binary images
• Volume Metrics (DICE, Coverage,ETC)

Requirements

The following non-standard libraries are required to use the full functionality of the project.

• SimpleITK
• PyRadiomics
• PyDicom
• Untangle
• numpy

Usage

Data Preparation

The data preparation step depends on the input data to be used.

Single Dataset

tpdo

Batch Processing

todo

python package_data.py --intraop_patch intra_op_patch.ply  --preop_ct pre_op.ply --intraop_ct intra_op.ply --output_file packaged.pckl

Patient Data

The data consists of a segmented pre-operative CT model and tracked images from a stereo-endoscope. The data has to be organized as follows:

.
├── ...
├── path_to_data            # input data folder
│   ├── *.jpg               # interlaced stereo images
│   ├── *.xml               # CASone xml parameter file
│   └── mask                # mask folder (optional)
│       ├── *.jpg           # segmentation masks with same name as stereo images
└── ...

Compute Radiomics

python compute_stereo.py --input_dir path_to_data --result_dir path_to_output --segment true

• to be completed
• to be completed
• to be completed

Plot

For experiments with synthetic data run (add optional parameters)

Output to CSV

python gp_experiments.py