A library that integrates different MIL methods into a unified framework

2024-07-20

• fix bug of all MIL-models expect DTFD-MIL

2024-07-21

• fix bug of DTFD-MIL
• fix bug of test_mil.py

2024-07-27

• fix bug of plip-transforms

• A library that integrates different MIL methods into a unified framework
• A library that integrates different Datasets into a unified Interface
• A library that provides different Datasets-Split-Methods which commonly used
• A library that easily extend by following a uniform definition

• User only need to provide the following csvs whether Public/Private Dataset
  /datasets/example_Dataset.csv

• User-difined Train-Val-Test split
• Train-Val split with K-fold
• Train-Val-Test split with K-fold

• R50 Deep Residual Learning for Image Recognition (CVPR 2016)
• VIT-S An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale (ICLR 2021)
• CTRANSPATH Transformer-Based Self-supervised Learning for Histopathological Image Classification (MIA 2023)
• PLIP A visual–language foundation model for pathology image analysis using medical Twitter (NAT MED 2023)
• CONCH A visual-language foundation model for computational pathology (NAT MED 2024)
• UNI Towards a general-purpose foundation model for computational pathology (NAT MED 2024)
• GIG A whole-slide foundation model for digital pathology from real-world data (NAT 2024)

• MEAN_MIL
• MAX_MIL
• AB_MIL Attention-based Deep Multiple Instance Learning (ICML 2018)
• TRANS_MIL Transformer based Correlated Multiple Instance Learning for WSI Classification (NeurIPS 2021)
• DS_MIL Dual-stream MIL Network for WSI Classification with Self-supervised Contrastive Learning (CVPR 2021)
• CLAM_MIL Data Efficient and Weakly Supervised Computational Pathology on WSI (NAT BIOMED ENG 2021)
• DTFD_MIL Double-Tier Feature Distillation MIL for Histopathology WSI Classification (CVPR 2022)
• RRT_MIL Towards Foundation Model-Level Performance in Computational Pathology (CVPR 2024)
• WIKG_MIL Dynamic Graph Representation with Knowledge-aware Attention for WSI Analysis (CVPR 2024)
• SC_MIL Sparse Context-aware MIL for Predicting Cancer Survival Probability (MICCAI 2024) UD...
• RET_MIL Retentive Multiple Instance Learning for Histopathological WSI Classification (MICCAI 2024) UD...
• UPDATING...

• AUC: macro,micro,weighed (same when 2-classes)
• F1,PRE,RECALL: macro,micro,weighed
• ACC,BACC: BACC is macro-RECALL
• KAPPLE: linear,quadratic
• Confusion_Mat
• FROC
• Patch-level-metrics

• /configs: MIL_BASELINE defines MIL models through a YAML configuration file.
• /modules: Defined the network architectures of different MIL models.
• /process: Defined the training frameworks for different MIL models.
• /feature_extracter: Supports different feature extractors.
• /datasets: User-Datasets path information.
• /utils: Framework's utility scripts.
• /train_mil.py: Train Entry function of the framework.
• /test_mil.py: Test Entry function of the framework.

• OpenSlide supported formats and SDPC formats
• R50 and VIT-S are supported directly.
• PLIP/UNI/COUCH/TRANSPATH/GIG are supported by push the model-weights in /feature_extracter/PLIP /feature_extracter/UNI /feature_extracter/COUCH /feature_extracter/CTRANSPATH /feature_extracter/GIG.
• To permform feature extracter
  
  • First, you should get h5 file
    python /feature_extracter/CLAM/create_patches_fp.py --source /path/to/your/slide_dir --save_dir /path/to/your/save_dr --preset /feature_extractor/CLAM/presets/bwh_biopsy.csv --patch_level your_path_level --patch_size your_patch_size --seg --patch
  • Second, you should get pt file
    CUDA_VISIBLE_DEVICES=your_cuda_id python /feature_extractor/CLAM/extract_features_fp.py --data_h5_dir /path/to/your/h5_save_dir --data_slide_dir /path/to/your/slide_dir --csv_path /path/to/your/h5_save_dir/process_list_autogen.csv --feat_dir path/to/your/pt_save_dir --batch_size your_bach_size --slide_ext your_ext --backbone your_backbone --target_patch_size your_target_patch_size --model_dir your_backbone_dir
    • backbone : resnet50_imagenet/vit_s_imagenet/plip/uni
    • ext : .svs/.tif/.ndpi/......./.sdpc
    • target_patch_size : for example, vit_s need 224 as input
    • ctranspath needs timm== timm-0.5.4 / gig needs timm=>1.0.3 / uni,conch... need timm==0.9.16

• You should construct a csv-file like the format of /datasets/example_Dataset.csv

• You can use the dataset-split-scripts to perform different data-split.

• you can config the yaml-file in /configs
• for example, /configs/AB_MIL.yaml, A detailed explanation has been written in /configs/AB_MIL.yaml
• Then, /train_mil.py will help you !
• /test_mil.py will help you test pretrained model !

• Personal experience is limited, and code submissions are welcome !!