Merge branch 'Project-MONAI:main' into maisi

Author: Can-Zhao

.pre-commit-config.yaml

@@ -23,7 +23,7 @@ repos:
       - id: detect-private-key
 
   - repo: https://github.com/psf/black
-    rev: "24.10.0"
+    rev: "25.1.0"
     hooks:
       - id: black
       - id: black-jupyter

2d_regression/image_restoration.ipynb

@@ -1,7 +1,18 @@
 3D Regression
 =============
 
-How to run the 3D regression tutorial.
---------------------------------------
+This directory contains a tutorial demonstrating how to use MONAI for 3D regression tasks, specifically brain age prediction using the IXI dataset and a DenseNet3D architecture.
 
-Running this notebook is straightforward.  It works well in Colab.
+## Tutorial Overview
+
+The `densenet_training_array.ipynb` notebook provides an end-to-end example of:
+- Loading and preprocessing 3D brain MRI data
+- Setting up data transforms for regression tasks
+- Training a DenseNet3D model for age prediction
+- Evaluating model performance on test data
+
+## How to Run
+
+This notebook can be run locally with Jupyter or in Google Colab. The notebook includes all necessary setup instructions and dependency installations.
+
+[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Project-MONAI/tutorials/blob/main/3d_regression/densenet_training_array.ipynb)

3d_regression/README.md

@@ -284,7 +284,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(keys=[\"image\", \"label\"], pixdim=(1.5, 1.5, 2.0), mode=(\"bilinear\", \"nearest\")),\n",
     "        RandCropByPosNegLabeld(\n",
@@ -318,7 +318,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(keys=[\"image\", \"label\"], pixdim=(1.5, 1.5, 2.0), mode=(\"bilinear\", \"nearest\")),\n",
     "    ]\n",
@@ -690,7 +690,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\"], source_key=\"image\", allow_smaller=True),\n",
     "    ]\n",
     ")\n",
     "\n",
@@ -784,7 +784,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\"], source_key=\"image\", allow_smaller=True),\n",
     "    ]\n",
     ")\n",
     "\n",

3d_segmentation/spleen_segmentation_3d.ipynb

@@ -278,7 +278,7 @@
     "                    b_max=1.0,\n",
     "                    clip=True,\n",
     "                ),\n",
-    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "                # randomly crop out patch samples from\n",
     "                # big image based on pos / neg ratio\n",
     "                # the image centers of negative samples\n",
@@ -321,7 +321,7 @@
     "                    b_max=1.0,\n",
     "                    clip=True,\n",
     "                ),\n",
-    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "            ]\n",
     "        )\n",
     "\n",

3d_segmentation/spleen_segmentation_3d_lightning.ipynb

@@ -317,7 +317,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(keys=[\"image\", \"label\"], pixdim=(1.5, 1.5, 2.0), mode=(\"bilinear\", \"nearest\")),\n",
     "        RandCropByPosNegLabeld(\n",
@@ -351,7 +351,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(keys=[\"image\", \"label\"], pixdim=(1.5, 1.5, 2.0), mode=(\"bilinear\", \"nearest\")),\n",
     "    ]\n",
@@ -782,7 +782,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\"], source_key=\"image\", allow_smaller=True),\n",
     "    ]\n",
     ")\n",
     "\n",

3d_segmentation/spleen_segmentation_3d_visualization_basic.ipynb

@@ -326,6 +326,7 @@
     "                keys=[\"image\", \"label\"],\n",
     "                source_key=\"image\",\n",
     "                k_divisible=[roi[0], roi[1], roi[2]],\n",
+    "                allow_smaller=True,\n",
     "            ),\n",
     "            transforms.RandSpatialCropd(\n",
     "                keys=[\"image\", \"label\"],\n",
@@ -454,7 +455,7 @@
    "source": [
     "## Create Swin UNETR model\n",
     "\n",
-    "In this scetion, we create Swin UNETR model for the 3-class brain tumor semantic segmentation. We use a feature size of 48. We also use gradient checkpointing (use_checkpoint) for more memory-efficient training. However, use_checkpoint for faster training if enough GPU memory is available.  "
+    "In this section, we create Swin UNETR model for the 3-class brain tumor semantic segmentation. We use a feature size of 48. We also use gradient checkpointing (use_checkpoint) for more memory-efficient training. However, use_checkpoint for faster training if enough GPU memory is available.  "
    ]
   },
   {
@@ -467,7 +468,6 @@
     "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
     "\n",
     "model = SwinUNETR(\n",
-    "    img_size=roi,\n",
     "    in_channels=4,\n",
     "    out_channels=3,\n",
     "    feature_size=48,\n",

3d_segmentation/swin_unetr_brats21_segmentation_3d.ipynb

@@ -96,7 +96,7 @@
    "source": [
     "# Pre-trained Swin UNETR Encoder\n",
     "\n",
-    "We use weights from self-supervised pre-training of Swin UNETR encoder (3D Swin Tranformer) on a cohort of 5050 CT scans from publicly available datasets. The encoder is pre-trained using reconstructin, rotation prediction and contrastive learning pre-text tasks as shown below. For more details, please refer to [1] (CVPR paper) and see this [repository](https://github.com/Project-MONAI/research-contributions/tree/main/SwinUNETR/Pretrain). \n",
+    "We use weights from self-supervised pre-training of Swin UNETR encoder (3D Swin Transformer) on a cohort of 5050 CT scans from publicly available datasets. The encoder is pre-trained using reconstructin, rotation prediction and contrastive learning pre-text tasks as shown below. For more details, please refer to [1] (CVPR paper) and see this [repository](https://github.com/Project-MONAI/research-contributions/tree/main/SwinUNETR/Pretrain). \n",
     "\n",
     "Please download the pre-trained weights from this [link](https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/model_swinvit.pt) and place it in the root directory of this tutorial. \n",
     "\n",
@@ -243,7 +243,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(\n",
     "            keys=[\"image\", \"label\"],\n",
@@ -292,7 +292,7 @@
     "    [\n",
     "        LoadImaged(keys=[\"image\", \"label\"], ensure_channel_first=True),\n",
     "        ScaleIntensityRanged(keys=[\"image\"], a_min=-175, a_max=250, b_min=0.0, b_max=1.0, clip=True),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "        Spacingd(\n",
     "            keys=[\"image\", \"label\"],\n",
@@ -439,7 +439,6 @@
     "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
     "\n",
     "model = SwinUNETR(\n",
-    "    img_size=(96, 96, 96),\n",
     "    in_channels=1,\n",
     "    out_channels=14,\n",
     "    feature_size=48,\n",
@@ -453,7 +452,7 @@
    "source": [
     "### Initialize Swin UNETR encoder from self-supervised pre-trained weights\n",
     "\n",
-    "In this section, we intialize the Swin UNETR encoder from pre-trained weights. The weights can be downloaded using the wget command below, or by following [this link](https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/model_swinvit.pt) to GitHub. If training from scratch is desired, please skip this section."
+    "In this section, we initialize the Swin UNETR encoder from pre-trained weights. The weights can be downloaded using the wget command below, or by following [this link](https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/model_swinvit.pt) to GitHub. If training from scratch is desired, please skip this section."
    ]
   },
   {

3d_segmentation/swin_unetr_btcv_segmentation_3d.ipynb

@@ -228,7 +228,7 @@
     "            b_max=1.0,\n",
     "            clip=True,\n",
     "        ),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "        RandCropByPosNegLabeld(\n",
     "            keys=[\"image\", \"label\"],\n",
     "            label_key=\"label\",\n",
@@ -277,7 +277,7 @@
     "            mode=(\"bilinear\", \"nearest\"),\n",
     "        ),\n",
     "        ScaleIntensityRanged(keys=[\"image\"], a_min=-175, a_max=250, b_min=0.0, b_max=1.0, clip=True),\n",
-    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "        CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "    ]\n",
     ")"
    ]

3d_segmentation/unetr_btcv_segmentation_3d.ipynb

@@ -466,7 +466,7 @@
     "                    b_max=1.0,\n",
     "                    clip=True,\n",
     "                ),\n",
-    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "                Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "                Spacingd(\n",
     "                    keys=[\"image\", \"label\"],\n",
@@ -522,7 +522,7 @@
     "                    b_max=1.0,\n",
     "                    clip=True,\n",
     "                ),\n",
-    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\"),\n",
+    "                CropForegroundd(keys=[\"image\", \"label\"], source_key=\"image\", allow_smaller=True),\n",
     "                Orientationd(keys=[\"image\", \"label\"], axcodes=\"RAS\"),\n",
     "                Spacingd(\n",
     "                    keys=[\"image\", \"label\"],\n",