RCC_DS

Code for paper "Renal Cell Carcinoma Detection and Subtyping with Minimal Point-Based Annotation in Whole-Slide Images". The dataset will be released soon. The original dataset comes from TCGA (KIRC, KIRP, KICH) project. https://portal.gdc.cancer.g

Pure Nim score 15/100 · last commit 2024-05-31 · 2 stars · tests present · no docs generated

Summary

Latest Version Unknown
License Unknown
CI Status Failing
Stars 2
Forks 1
Open Issues 0
Last Commit 2024-05-31
Downloads 0
Last Indexed 2026-08-11 05:07

Installation

nimble install RCC_DS
choosenim install RCC_DS
git clone https://gitlab.com/BioAI/RCC_DS

OS Compatibility

Platform Linux macOS Windows FreeBSD OpenBSD NetBSD Android iOS WASM Embedded
RCC_DS - - - - - - -

Source

Repository https://gitlab.com/BioAI/RCC_DS
Homepage https://gitlab.com/BioAI/RCC_DS
Registry Source gitlab

README

RCC Detection and Subtyping Framework

Code for paper "Renal Cell Carcinoma Detection and Subtyping with Minimal Point-Based Annotation in Whole-Slide Images".

The original dataset comes from TCGA (KIRC, KIRP, KICH) project.

The annotated dataset is released on https://dataset.chenli.group/home/rcc-region-and-subtyping.

Setups

The requirement is as bellow:

  • Python 3+
  • PyTorch 1.3.1
  • Torchvision 0.4.2
  • numpy 1.16.2
  • PIL 6.2.1
  • matplotlib 3.0.2
  • tqdm 4.28.1
  • sklearn 0.20.1
  • Openslide 1.1.1
  • cv2 4.1.2.30
  • pandas 0.23.4

Running

Detection (CCRCC example)

# train
python RCC_detection_train.py --epochs 200 --labeled_data_files your_labeled_images.txt --unlabeled_files your_unlabeled_images.txt --test_files valid_images.txt --out rcc@detection
# finetune
python RCC_detection_train.py --epochs 205 --resume rcc@detection/checkpoint.pth.tar --transfer True --labeled_data_files your_labeled_images.txt --unlabeled_files other_unlabeled_images.txt --test_files valid_images.txt --out rcc@detection@transfer
# predict
python RCC_predict.py --num_classes 2 --file_path_base rcc@detection@transfer/checkpoint.pth.tar --test_files your_unlabeled_images.txt --output_files your_unlabeled_predicted.txt
python RCC_predict.py --num_classes 2 --file_path_base rcc@detection@transfer/checkpoint.pth.tar --test_files other_unlabeled_images.txt --output_files other_unlabeled_predicted.txt
cat your_unlabeled_predicted.txt,other_unlabeled_predicted.txt -> ccrcc_predicted.txt

Subtyping

# train
cat ccrcc_predicted.txt,prcc_predicted.txt,chrcc_predicted.txt -> subtype_train_images.txt
python RCC_subtyping_train.py --num_classes 4 --model_path_base ./subtype_model/checkpoint.pth --model_path_best ./subtype_model/model_best.pth --train_files subtype_train_images.txt --test_files subtype_valid_images.txt 
# predict
python RCC_predict.py --num_classes 4 --file_path_base ./subtype_model/model_best.pth --test_files subtype_test_images.txt