Implement Feature/tiger-wsibulk

configs/vision/pathology/offline/classification/tiger_tumour.yaml

@@ -0,0 +1,134 @@
+---
+trainer:
+  class_path: eva.Trainer
+  init_args:
+    n_runs: &N_RUNS ${oc.env:N_RUNS, 20}
+    default_root_dir: &OUTPUT_ROOT ${oc.env:OUTPUT_ROOT, logs/${oc.env:MODEL_NAME, dino_vits16}/offline/tiger_tumour}
+    max_epochs: &MAX_EPOCHS ${oc.env:MAX_EPOCHS, 100}
+    checkpoint_type: ${oc.env:CHECKPOINT_TYPE, best}
+    callbacks:
+      - class_path: eva.callbacks.ConfigurationLogger
+      - class_path: lightning.pytorch.callbacks.TQDMProgressBar
+        init_args:
+          refresh_rate: ${oc.env:TQDM_REFRESH_RATE, 1}
+      - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+        init_args:
+          logging_interval: epoch
+      - class_path: lightning.pytorch.callbacks.ModelCheckpoint
+        init_args:
+          filename: best
+          save_last: ${oc.env:SAVE_LAST, false}
+          save_top_k: 1
+          monitor: &MONITOR_METRIC ${oc.env:MONITOR_METRIC, val/BinaryBalancedAccuracy}
+          mode: &MONITOR_METRIC_MODE ${oc.env:MONITOR_METRIC_MODE, max}
+      - class_path: lightning.pytorch.callbacks.EarlyStopping
+        init_args:
+          min_delta: 0
+          patience: ${oc.env:PATIENCE, 20}
+          monitor: *MONITOR_METRIC
+          mode: *MONITOR_METRIC_MODE
+      - class_path: eva.callbacks.ClassificationEmbeddingsWriter
+        init_args:
+          output_dir: &DATASET_EMBEDDINGS_ROOT ${oc.env:EMBEDDINGS_ROOT, ./data/embeddings/${oc.env:MODEL_NAME, dino_vits16}/tiger_tumour}
+          dataloader_idx_map:
+            0: train
+            1: val
+            2: test
+          metadata_keys: ["wsi_id"]
+          backbone:
+            class_path: eva.vision.models.ModelFromRegistry
+            init_args:
+              model_name: ${oc.env:MODEL_NAME, universal/vit_small_patch16_224_dino}
+              model_extra_kwargs: ${oc.env:MODEL_EXTRA_KWARGS, null}
+          overwrite: false
+    logger:
+      - class_path: lightning.pytorch.loggers.TensorBoardLogger
+        init_args:
+          save_dir: *OUTPUT_ROOT
+          name: ""
+model:
+  class_path: eva.HeadModule
+  init_args:
+    head:
+      class_path: eva.vision.models.networks.ABMIL
+      init_args:
+        input_size: ${oc.env:IN_FEATURES, 384}
+        output_size: &NUM_CLASSES 1
+        projected_input_size: 128
+    criterion: torch.nn.BCEWithLogitsLoss
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: ${oc.env:LR_VALUE, 0.001}
+        betas: [0.9, 0.999]
+    metrics:
+      common:
+        - class_path: eva.metrics.AverageLoss
+        - class_path: eva.metrics.BinaryClassificationMetrics
+data:
+  class_path: eva.DataModule
+  init_args:
+    datasets:
+      train:
+        class_path: eva.datasets.MultiEmbeddingsClassificationDataset
+        init_args: &DATASET_ARGS
+          root: *DATASET_EMBEDDINGS_ROOT
+          manifest_file: manifest.csv
+          split: train
+          embeddings_transforms:
+            class_path: eva.core.data.transforms.Pad2DTensor
+            init_args:
+              pad_size: &N_PATCHES ${oc.env:N_PATCHES, 200}
+          target_transforms:
+            class_path: eva.core.data.transforms.dtype.ArrayToFloatTensor
+      val:
+        class_path: eva.datasets.MultiEmbeddingsClassificationDataset
+        init_args:
+          <<: *DATASET_ARGS
+          split: val
+      test:
+        class_path: eva.datasets.MultiEmbeddingsClassificationDataset
+        init_args:
+          <<: *DATASET_ARGS
+          split: test
+      predict:
+        - class_path: eva.vision.datasets.TIGERTumour
+          init_args: &PREDICT_DATASET_ARGS
+            root: ${oc.env:DATA_ROOT, ./data/training/wsibulk}
+            sampler:
+              class_path: eva.vision.data.wsi.patching.samplers.ForegroundGridSampler
+              init_args:
+                max_samples: *N_PATCHES
+            width: 224
+            height: 224
+            target_mpp: 0.5
+            split: train
+            coords_path: ${data.init_args.datasets.train.init_args.root}/coords_${.split}.csv
+            image_transforms:
+              class_path: eva.vision.data.transforms.common.ResizeAndCrop
+              init_args:
+                size: ${oc.env:RESIZE_DIM, 224}
+                mean: ${oc.env:NORMALIZE_MEAN, [0.485, 0.456, 0.406]}
+                std: ${oc.env:NORMALIZE_STD, [0.229, 0.224, 0.225]}
+        - class_path: eva.vision.datasets.TIGERTumour
+          init_args:
+            <<: *PREDICT_DATASET_ARGS
+            split: val
+        - class_path: eva.vision.datasets.TIGERTumour
+          init_args:
+            <<: *PREDICT_DATASET_ARGS
+            split: test
+    dataloaders:
+      train:
+        batch_size: &BATCH_SIZE ${oc.env:BATCH_SIZE, 32}
+        num_workers: &N_DATA_WORKERS ${oc.env:N_DATA_WORKERS, 4}
+        shuffle: true
+      val:
+        batch_size: *BATCH_SIZE
+        num_workers: *N_DATA_WORKERS
+      test:
+        batch_size: *BATCH_SIZE
+        num_workers: *N_DATA_WORKERS
+      predict:
+        batch_size: &PREDICT_BATCH_SIZE ${oc.env:PREDICT_BATCH_SIZE, 64}
+        num_workers: *N_DATA_WORKERS

docs/datasets/tiger.md

@@ -0,0 +1,69 @@
+# TIGER (Tumor Infiltrating Lymphocytes in breast cancER)
+
+TIGER contains digital pathology images of Her2 positive (Her2+) and Triple Negative (TNBC) breast cancer whole-slide images, together with manual annotations. Training data comes from multiple sources. A subset of Her2+ and TNBC cases is provided by the Radboud University Medical Center (RUMC) (Nijmegen, Netherlands). A second subset of Her2+ and TNBC cases is provided by the Jules Bordet Institut (JB) (Bruxelles, Belgium). A third subset of TNBC cases only is derived from the TCGA-BRCA archive obtained from the Genomic Data Commons Data Portal.
+
+It contains 3 different datasets and thus 3 different tasks to add to eva. However only two are currently added.
+
+WSIBULK - WSI level classification task: Detecting tumour presence in patches of a given slide.
+WSITILS - Regression task: predicting "TIL" score of a whole slide image. 
+
+Source: https://tiger.grand-challenge.org/Data/ 
+
+
+## Raw data
+
+### Key stats
+
+|                           |                                                          |
+|---------------------------|----------------------------------------------------------|
+| **Modality**              | Vision (WSI)                                             |
+| **Tasks**                 | Binary Classification / Regression                       |
+| **Cancer type**           | Breast                                                   |
+| **Data size**             | 182 GB                                                   |
+| **Image dimension**       | ~20k x 20k x 3                                           |
+| **Magnification (μm/px)** | 20x (0.5) - Level 0                                      |
+| **Files format**          | `.tif`                                                   |
+| **Number of images**      | 178 WSIs (96 for WSIBULK and 82 for WSITILS)             |
+
+
+### Organization
+
+The data `tiger.zip` from [grand challenge](https://tiger.grand-challenge.org/) is organized as follows:
+
+training/
+	|_wsibulk/ (used for classification task)  
+	|	|__annotations-tumor-bulk/				* manual annotations of "tumor bulk" regions (see https://tiger.grand-challenge.org/Data/ for details)
+	|	|	|___masks/							* annotations in multiresolution TIF format
+	|	|	|___xmls/ (not used in eva)
+	|	|__images/	
+
+	|	|__tissue-masks/ (not used in eva)		
+	|
+	|_wsirois/ (not used in eva yet)
+	|
+	|_wsitils/	(used for regression task)							
+	|	|__images/	
+    |   │   ├── 104S.tiff
+    │   |   └── ...									* whole-slide images
+	|	|__tissue-masks/ (not used in eva)	
+	|	|__tiger-tils-scores-wsitils.csv (target variable file)
+
+
+
+
+## Download and preprocessing
+
+The `TIGER` dataset class doesn't download the data during runtime and must be downloaded manually as follows:
+
+- Make sure that the latest version of the AWS CLI is installed on your system by following [these instructions](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html)
+
+With the AWS CLI installed, you can download the public training set (no AWS account required) by running:
+
+`aws s3 cp s3://tiger-training/ /path/to/destination/ --recursive --no-sign-request`
+
+
+We then generate random stratified train / validation and test splits using a 0.7 / 0.15 / 0.15 ratio.
+
+
+
+

src/eva/vision/data/datasets/__init__.py

@@ -11,6 +11,7 @@
     GleasonArvaniti,
     PANDASmall,
     PatchCamelyon,
+    TIGERTumour,
     UniToPatho,
     WsiClassificationDataset,
 )
@@ -49,4 +50,5 @@
     "VisionDataset",
     "MultiWsiDataset",
     "WsiDataset",
+    "TIGERTumour",
 ]

src/eva/vision/data/datasets/classification/__init__.py

@@ -9,6 +9,7 @@
 from eva.vision.data.datasets.classification.mhist import MHIST
 from eva.vision.data.datasets.classification.panda import PANDA, PANDASmall
 from eva.vision.data.datasets.classification.patch_camelyon import PatchCamelyon
+from eva.vision.data.datasets.classification.tiger_tumour import TIGERTumour
 from eva.vision.data.datasets.classification.unitopatho import UniToPatho
 from eva.vision.data.datasets.classification.wsi import WsiClassificationDataset
 
@@ -26,4 +27,5 @@
     "PANDA",
     "PANDASmall",
     "Camelyon16",
+    "TIGERTumour",
 ]