\n",
+ "```\n",
+ "\n",
+ "**Step 2**: Start QuPath with excutable launcher file.\n",
+ "\n",
+ "Start QuPath by opening a **Terminal**, Use ```chmod u+x /path/to/QuPath/bin/QuPath``` to make the extracted QuPath launcher executable.\n",
+ "If you suffer from problem, please refer to QuPath [QA](https://github.com/qupath/qupath/releases/tag/v0.3.2).\n",
+ "\n",
+ "Then type or drag the launcher file: ```/path/to/QuPath/bin/QuPath``` to the terminal, and enter to open QuPath UI.\n",
+ "\n",
+ "**Step 3** Install MONAI Label Plugin for QuPath\n",
+ "\n",
+ "Download [qupath-extension-monailabel-0.3.1.jar](https://github.com/Project-MONAI/MONAILabel/releases/download/data/qupath-extension-monailabel-0.3.1.jar)\n",
+ "and **drag the jar** on top of the running QuPath application window (black screen area) to install the extension.\n",
+ "If you have previously installed then make sure to **_remove/uninstall_** the extension before updating.\n",
+ "\n",
+ "![\"drawing\"](\"./figures/monailabel_pathology_qupath/qupath_plugin.png\")
\n",
+ "Drag the downloaded MONAI Label QuPath plugin JAR file into QuPath viewer UI.
\n",
+ "\n",
+ "Note: Make sure MONAILabel Server URL (Host+Port) is correct through `Preferences`.\n",
+ "\n"
+ ]
+ },
+ {
+ "attachments": {},
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## 3. Nuclei Classification and Segmentation with Interactions\n",
+ "\n",
+ "Below is the instruction of using HoVerNet for labeling nucleis given WSI or other digital pathology images using QuPath. Pre-trained model is provided and loaded by MONAI Label server.\n",
+ "\n",
+ "\n",
+ "### 3.1 Start QuPath, MONAI Label Plugin, and Fetch Next Image\n",
+ "\n",
+ "- On the menu bar, navigate and click MONAI Label. Click *Next Sample/Patch* to load data specified in the start_server command. If the QuPath starts from beginning, the button will automatically load next whole slide image.\n",
+ "\n",
+ "![\"drawing\"](\"./figures/monailabel_pathology_qupath/qupath2.png\")
\n",
+ "\n",
+ " - Users can also use the **Next Sample/Patch** button to select a patch ROI automatically, by selecting **Next Patch**. （Uncheck the **Next Patch** box will load next whole slide image.）\n",
+ "\n",
+ "\n",
+ "\n",
+ "### 3.2 Select ROI by Rectangle Tool\n",
+ "\n",
+ "First, let's show patch inference with the rectangle ROI tool.\n",
+ "\n",
+ "- Users can automatically select the rectangle ROI by fetching \"Next Patch\", or use the QuPath rectangle tool to pick an ROI.\n",
+ "\n",
+ "\n",
+ "\n",
+ "### 3.3 Run Nuclei Classication and Segmentation Simultaneously\n",
+ "\n",
+ "- After selecting the rectangle ROI, users can navigate to the MONAI Label plugin and click **Annotations**.\n",
+ "Then select models in the pop-up interface, and click **OK** to run inference. \n",
+ "\n",
+ "\n",
+ "\n",
+ "MONAI Label server will run auto segmentation with the selected patch data and return the segmentation with polygons.\n",
+ "\n",
+ "### 3.4 Submit ROI Label for Training\n",
+ "\n",
+ "- In the next step, users can edit the predicted nuclei segmentation. Once the annotation is done, users can navigate to MONAI Label plugin and click **Submit Label** to save the ground truth.\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "### 3.5 Run Interactive ROI-based Segmentation\n",
+ "\n",
+ "- Users can also use the interactive ROI tool to get auto-segmentation with one ROI draw:\n",
+ "- Draw any rectangle ROI, and MONAI Label server will return segmentation immediately:\n",
+ "\n",
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## 4. Training Nuclei Segmentation\n",
+ "\n",
+ "Once users finish annotations on several patch images, MONAI Label provide training option to fine-tune the model.\n",
+ "\n",
+ "- Navigate to MONAI Label plugin and click **Training**\n",
+ "\n",
+ "\n",
+ "\n",
+ "- In the training option pop-up interface, select **hovernet_nuclei**, select training customized training options. Uncheck **pretrained** to start training model from scratch. Finally, click **OK** to start training in MONAI Label server.\n",
+ "\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## 5. Repeat the Interactive Labeling Process \n",
+ "\n",
+ "Repeated fetch data and interactive learning process to section 3.1, until a satisfied model is achieved or entire datasets/WSIs are annotated. \n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## 6. Conclusion\n",
+ "\n",
+ "This tutorial demonstrates the basic usage of HoVerNet, nuclei classification and segmentation at the same time with MONAI Label. We introduced several inference strategies including ROI tool, interactive ROI tool. We show how to submit annotated labels, then the steps of training are demonstrated.\n",
+ "\n",
+ "In the MONAI Label HoVerNet integration, we highlight the interactive labeling experience with classification and segmentation at the same time from a whole slide image. MONAI Label achieves annotation actively and gradually improves the model's classification and segmentation performance during labeling. \n"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.8.10"
+ },
+ "vscode": {
+ "interpreter": {
+ "hash": "916dbcbb3f70747c44a77c7bcd40155683ae19c65e1c03b4aa3499c5328201f1"
+ }
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/runner.sh b/runner.sh
index 7d2db4e3f6..449efd85f5 100755
--- a/runner.sh
+++ b/runner.sh
@@ -58,6 +58,7 @@ doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" monailabel_patholog
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" monailabel_radiology_spleen_segmentation_OHIF.ipynb)
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" monailabel_wholebody_totalSegmentator_3dslicer.ipynb)
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" monailabel_monaibundle_3dslicer_lung_nodule_detection.ipynb)
+doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" monailabel_pathology_HoVerNet_QuPath.ipynb)
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" example_feature.ipynb)
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" ssl_train.ipynb)
doesnt_contain_max_epochs=("${doesnt_contain_max_epochs[@]}" ssl_finetune.ipynb)
@@ -93,6 +94,7 @@ skip_run_papermill=("${skip_run_papermill[@]}" .*monailabel_pathology_nuclei_seg
skip_run_papermill=("${skip_run_papermill[@]}" .*monailabel_radiology_spleen_segmentation_OHIF*)
skip_run_papermill=("${skip_run_papermill[@]}" .*monailabel_wholebody_totalSegmentator_3dslicer*)
skip_run_papermill=("${skip_run_papermill[@]}" .*monailabel_monaibundle_3dslicer_lung_nodule_detection*)
+skip_run_papermill=("${skip_run_papermill[@]}" .*monailabel_pathology_HoVerNet_QuPath*)
skip_run_papermill=("${skip_run_papermill[@]}" .*ssl_train*)
skip_run_papermill=("${skip_run_papermill[@]}" .*ssl_finetune*)
skip_run_papermill=("${skip_run_papermill[@]}" .*transform_visualization*) # https://github.com/Project-MONAI/tutorials/issues/1155