PCa classification optimisation experiments based on the radiomics features - simple models that base on the sklearn classifiers. See: https://www.researchgate.net/publication/317660491_MRI_imaging_texture_features_in_prostate_lesions_classification https://www.researchgate.net/publication/317721536_Feature_Extraction_Optimized_For_Prostate_Lesion_Classification
Python scripts with experiments base on the corresponding .json configuration files (see example for very brief documentation)
- Download ProstateX data
- Enforce correct directory structure
/{project.dir}/data/ProstateX/train/DOI
/{project.dir}/data/ProstateX/train/ktrans
/{project.dir}/data/ProstateX/train/lesion-information/
/{project.dir}/data/ProstateX/train/lesion-information/
/{project.dir}/data/ProstateX/train/lesion-information/ProstateX.csv (if absent generated automatically by Dataset) - a dataset file containing data merged from patients mhd files and meta-data files (csv)
/{project.dir}/data/ProstateX/train/lesion-information/ProstateX-DataInfo-Test.docx
/{project.dir}/data/ProstateX/train/lesion-information/ProstateX-Findings.csv
/{project.dir}/data/ProstateX/train/lesion-information/ProstateX-Images.csv
/{project.dir}/data/ProstateX/train/lesion-information/ProstateX-Images-KTrans.csv
/{project.dir}/data/ProstateX/train/screenshots (optional)