complete the model package

LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

README.md

@@ -0,0 +1,154 @@
+---
+tags:
+- monai
+- medical
+library_name: monai
+license: apache-2.0
+---
+# Model Overview
+
+A pre-trained model for simultaneous segmentation and classification of nuclei within multi-tissue histology images based on CoNSeP data. The details of the model can be found in [1].
+
+## Workflow
+
+The model is trained to simultaneous segment and classify nuclei. Training is done via a two-stage approach. First initialized the model with pre-trained weights on the [ImageNet dataset](https://ieeexplore.ieee.org/document/5206848), trained only the decoders for the first 50 epochs, and then fine-tuned all layers for another 50 epochs. There are two training modes in total. If "original" mode is specified, it uses [270, 270] and [80, 80] for `patch_size` and `out_size` respectively. If "fast" mode is specified, it uses [256, 256] and [164, 164] for `patch_size` and `out_size` respectively. The results we show below are based on the "fast" model.
+
+- We train the first stage with pre-trained weights from some internal data.
+
+- The original author's repo also has pre-trained weights which is for non-commercial use. Each user is responsible for checking the content of models/datasets and the applicable licenses and determining if suitable for the intended use. The license for the pre-trained model is different than MONAI license. Please check the source where these weights are obtained from: <https://github.com/vqdang/hover_net#data-format>
+
+`PRETRAIN_MODEL_URL` is "https://drive.google.com/u/1/uc?id=1KntZge40tAHgyXmHYVqZZ5d2p_4Qr2l5&export=download" which can be used in bash code below.
+
+![Model workflow](https://ars.els-cdn.com/content/image/1-s2.0-S1361841519301045-fx1_lrg.jpg)
+
+## Data
+
+The training data is from <https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/>.
+
+- Target: segment instance-level nuclei and classify the nuclei type
+- Task: Segmentation and classification
+- Modality: RGB images
+- Size: 41 image tiles (2009 patches)
+
+The provided labelled data was partitioned, based on the original split, into training (27 tiles) and testing (14 tiles) datasets.
+
+After download the datasets, please run `scripts/prepare_patches.py` to prepare patches from tiles. Prepared patches are saved in `your-concep-dataset-path`/Prepared. The implementation is referring to <https://github.com/vqdang/hover_net/blob/master/extract_patches.py>. The command is like:
+
+```
+python scripts/prepare_patches.py -root your-concep-dataset-path
+```
+
+## Training configuration
+
+This model utilized a two-stage approach. The training was performed with the following:
+
+- GPU: At least 24GB of GPU memory.
+- Actual Model Input: 256 x 256
+- AMP: True
+- Optimizer: Adam
+- Learning Rate: 1e-4
+- Loss: HoVerNetLoss
+
+## Input
+
+Input: RGB images
+
+## Output
+
+Output: a dictionary with the following keys:
+
+1. nucleus_prediction: predict whether or not a pixel belongs to the nuclei or background
+2. horizontal_vertical: predict the horizontal and vertical distances of nuclear pixels to their centres of mass
+3. type_prediction: predict the type of nucleus for each pixel
+
+## Model Performance
+
+The achieved metrics on the validation data are:
+
+Fast mode:
+- Binary Dice: 0.8293
+- PQ: 0.4936
+- F1d: 0.7480
+
+#### Training Loss and Dice
+
+stage1:
+![A graph showing the training loss and the mean dice over 50 epochs in stage1](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_train_stage1_fast.png)
+
+stage2:
+![A graph showing the training loss and the mean dice over 50 epochs in stage2](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_train_stage2_fast.png)
+
+#### Validation Dice
+
+stage1:
+
+![A graph showing the validation mean dice over 50 epochs in stage1](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_val_stage1_fast.png)
+
+stage2:
+
+![A graph showing the validation mean dice over 50 epochs in stage2](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_val_stage2_fast.png)
+
+## commands example
+
+Execute training:
+
+- Run first stage
+
+```
+python -m monai.bundle run training --meta_file configs/metadata.json --config_file configs/train.json --logging_file configs/logging.conf --network_def#pretrained_url `PRETRAIN_MODEL_URL` --stage 0
+```
+
+- Run second stage
+
+```
+python -m monai.bundle run training --meta_file configs/metadata.json --config_file configs/train.json --logging_file configs/logging.conf --network_def#freeze_encoder false --network_def#pretrained_url None --stage 1
+```
+
+Override the `train` config to execute multi-GPU training:
+
+- Run first stage
+
+```
+torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run training --meta_file configs/metadata.json --config_file "['configs/train.json','configs/multi_gpu_train.json']" --logging_file configs/logging.conf --train#dataloader#batch_size 8 --network_def#freeze_encoder true --network_def#pretrained_url `PRETRAIN_MODEL_URL` --stage 0
+```
+
+- Run second stage
+
+```
+torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run training --meta_file configs/metadata.json --config_file "['configs/train.json','configs/multi_gpu_train.json']" --logging_file configs/logging.conf --train#dataloader#batch_size 4 --network_def#freeze_encoder false --network_def#pretrained_url None --stage 1
+```
+
+Override the `train` config to execute evaluation with the trained model:
+
+```
+python -m monai.bundle run evaluating --meta_file configs/metadata.json --config_file "['configs/train.json','configs/evaluate.json']" --logging_file configs/logging.conf
+```
+
+### Execute inference
+
+```
+python -m monai.bundle run evaluating --meta_file configs/metadata.json --config_file configs/inference.json --logging_file configs/logging.conf
+```
+
+# Disclaimer
+
+This is an example, not to be used for diagnostic purposes.
+
+# References
+
+[1] Simon Graham, Quoc Dang Vu, Shan E Ahmed Raza, Ayesha Azam, Yee Wah Tsang, Jin Tae Kwak, Nasir Rajpoot, Hover-Net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images, Medical Image Analysis, 2019 https://doi.org/10.1016/j.media.2019.101563
+
+# License
+Copyright (c) MONAI Consortium
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

configs/evaluate.json

@@ -0,0 +1,38 @@
+{
+    "network_def": {
+        "_target_": "HoVerNet",
+        "mode": "@hovernet_mode",
+        "adapt_standard_resnet": true,
+        "in_channels": 3,
+        "out_classes": 5
+    },
+    "validate#handlers": [
+        {
+            "_target_": "CheckpointLoader",
+            "load_path": "$os.path.join(@bundle_root, 'models', 'model.pt')",
+            "load_dict": {
+                "model": "@network"
+            }
+        },
+        {
+            "_target_": "StatsHandler",
+            "iteration_log": false
+        },
+        {
+            "_target_": "MetricsSaver",
+            "save_dir": "@output_dir",
+            "metrics": [
+                "val_mean_dice"
+            ],
+            "metric_details": [
+                "val_mean_dice"
+            ],
+            "batch_transform": "$monai.handlers.from_engine(['image_meta_dict'])",
+            "summary_ops": "*"
+        }
+    ],
+    "evaluating": [
+        "$setattr(torch.backends.cudnn, 'benchmark', True)",
+        "$@validate#evaluator.run()"
+    ]
+}

configs/inference.json

@@ -0,0 +1,151 @@
+{
+    "imports": [
+        "$import glob",
+        "$import os"
+    ],
+    "bundle_root": "$os.getcwd()",
+    "output_dir": "$os.path.join(@bundle_root, 'eval')",
+    "dataset_dir": "/workspace/Data/Pathology/CoNSeP/Test/Images",
+    "num_cpus": 2,
+    "batch_size": 1,
+    "sw_batch_size": 16,
+    "hovernet_mode": "fast",
+    "patch_size": 256,
+    "out_size": 164,
+    "device": "$torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')",
+    "network_def": {
+        "_target_": "HoVerNet",
+        "mode": "@hovernet_mode",
+        "adapt_standard_resnet": true,
+        "in_channels": 3,
+        "out_classes": 5
+    },
+    "network": "$@network_def.to(@device)",
+    "preprocessing": {
+        "_target_": "Compose",
+        "transforms": [
+            {
+                "_target_": "LoadImaged",
+                "keys": "image",
+                "reader": "$monai.data.PILReader",
+                "converter": "$lambda x: x.convert('RGB')"
+            },
+            {
+                "_target_": "EnsureChannelFirstd",
+                "keys": "image"
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": "image",
+                "dtype": "float32"
+            },
+            {
+                "_target_": "ScaleIntensityRanged",
+                "keys": "image",
+                "a_min": 0.0,
+                "a_max": 255.0,
+                "b_min": 0.0,
+                "b_max": 1.0,
+                "clip": true
+            }
+        ]
+    },
+    "data_list": "$[{'image': image} for image in glob.glob(os.path.join(@dataset_dir, '*.png'))]",
+    "dataset": {
+        "_target_": "Dataset",
+        "data": "@data_list",
+        "transform": "@preprocessing"
+    },
+    "dataloader": {
+        "_target_": "DataLoader",
+        "dataset": "@dataset",
+        "batch_size": "@batch_size",
+        "shuffle": false,
+        "num_workers": "@num_cpus",
+        "pin_memory": true
+    },
+    "inferer": {
+        "_target_": "SlidingWindowHoVerNetInferer",
+        "roi_size": "@patch_size",
+        "sw_batch_size": "@sw_batch_size",
+        "overlap": "$1.0 - float(@out_size) / float(@patch_size)",
+        "padding_mode": "constant",
+        "cval": 0,
+        "progress": true,
+        "extra_input_padding": "$((@patch_size - @out_size) // 2,) * 4"
+    },
+    "postprocessing": {
+        "_target_": "Compose",
+        "transforms": [
+            {
+                "_target_": "FlattenSubKeysd",
+                "keys": "pred",
+                "sub_keys": [
+                    "horizontal_vertical",
+                    "nucleus_prediction",
+                    "type_prediction"
+                ],
+                "delete_keys": true
+            },
+            {
+                "_target_": "HoVerNetInstanceMapPostProcessingd",
+                "sobel_kernel_size": 21,
+                "marker_threshold": 0.4,
+                "marker_radius": 2
+            },
+            {
+                "_target_": "HoVerNetNuclearTypePostProcessingd"
+            },
+            {
+                "_target_": "FromMetaTensord",
+                "keys": [
+                    "image"
+                ]
+            },
+            {
+                "_target_": "SaveImaged",
+                "keys": "instance_map",
+                "meta_keys": "image_meta_dict",
+                "output_ext": ".nii.gz",
+                "output_dir": "@output_dir",
+                "output_postfix": "instance_map",
+                "output_dtype": "uint32",
+                "separate_folder": false
+            },
+            {
+                "_target_": "SaveImaged",
+                "keys": "type_map",
+                "meta_keys": "image_meta_dict",
+                "output_ext": ".nii.gz",
+                "output_dir": "@output_dir",
+                "output_postfix": "type_map",
+                "output_dtype": "uint8",
+                "separate_folder": false
+            }
+        ]
+    },
+    "handlers": [
+        {
+            "_target_": "CheckpointLoader",
+            "load_path": "$os.path.join(@bundle_root, 'models', 'model.pt')",
+            "map_location": "@device",
+            "load_dict": {
+                "model": "@network"
+            }
+        }
+    ],
+    "evaluator": {
+        "_target_": "SupervisedEvaluator",
+        "device": "@device",
+        "val_data_loader": "@dataloader",
+        "val_handlers": "@handlers",
+        "network": "@network",
+        "postprocessing": "@postprocessing",
+        "inferer": "@inferer",
+        "amp": true
+    },
+    "evaluating": [
+        "$setattr(torch.backends.cudnn, 'benchmark', True)",
+        "[email protected]()"
+    ]
+}

configs/logging.conf

@@ -0,0 +1,21 @@
+[loggers]
+keys=root
+
+[handlers]
+keys=consoleHandler
+
+[formatters]
+keys=fullFormatter
+
+[logger_root]
+level=INFO
+handlers=consoleHandler
+
+[handler_consoleHandler]
+class=StreamHandler
+level=INFO
+formatter=fullFormatter
+args=(sys.stdout,)
+
+[formatter_fullFormatter]
+format=%(asctime)s - %(name)s - %(levelname)s - %(message)s

configs/metadata.json

@@ -0,0 +1,115 @@
+{
+    "schema": "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/meta_schema_hovernet_20221124.json",
+    "version": "0.1.0",
+    "changelog": {
+        "0.1.0": "complete the model package"
+    },
+    "monai_version": "1.1.0rc2",
+    "pytorch_version": "1.13.0",
+    "numpy_version": "1.22.2",
+    "optional_packages_version": {
+        "scikit-image": "0.19.3",
+        "scipy": "1.8.1",
+        "tqdm": "4.64.1",
+        "pillow": "9.0.1"
+    },
+    "task": "Nuclear segmentation and classification",
+    "description": "A simultaneous segmentation and classification of nuclei within multitissue histology images based on CoNSeP data",
+    "authors": "MONAI team",
+    "copyright": "Copyright (c) MONAI Consortium",
+    "data_source": "https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/",
+    "data_type": "numpy",
+    "image_classes": "RGB image with intensity between 0 and 255",
+    "label_classes": "a dictionary contains binary nuclear segmentation, hover map and pixel-level classification",
+    "pred_classes": "a dictionary contains scalar probability for binary nuclear segmentation, hover map and pixel-level classification",
+    "eval_metrics": {
+        "Binary Dice": 0.8293,
+        "PQ": 0.4936,
+        "F1d": 0.748
+    },
+    "intended_use": "This is an example, not to be used for diagnostic purposes",
+    "references": [
+        "Simon Graham. 'HoVer-Net: Simultaneous Segmentation and Classification of Nuclei in Multi-Tissue Histology Images.' Medical Image Analysis, 2019. https://arxiv.org/abs/1812.06499"
+    ],
+    "network_data_format": {
+        "inputs": {
+            "image": {
+                "type": "image",
+                "format": "magnitude",
+                "num_channels": 3,
+                "spatial_shape": [
+                    "256",
+                    "256"
+                ],
+                "dtype": "float32",
+                "value_range": [
+                    0,
+                    255
+                ],
+                "is_patch_data": true,
+                "channel_def": {
+                    "0": "image"
+                }
+            }
+        },
+        "outputs": {
+            "nucleus_prediction": {
+                "type": "probability",
+                "format": "segmentation",
+                "num_channels": 3,
+                "spatial_shape": [
+                    "164",
+                    "164"
+                ],
+                "dtype": "float32",
+                "value_range": [
+                    0,
+                    1
+                ],
+                "is_patch_data": true,
+                "channel_def": {
+                    "0": "background",
+                    "1": "nuclei"
+                }
+            },
+            "horizontal_vertical": {
+                "type": "probability",
+                "format": "regression",
+                "num_channels": 2,
+                "spatial_shape": [
+                    "164",
+                    "164"
+                ],
+                "dtype": "float32",
+                "value_range": [
+                    0,
+                    1
+                ],
+                "is_patch_data": true,
+                "channel_def": {
+                    "0": "horizontal distances map",
+                    "1": "vertical distances map"
+                }
+            },
+            "type_prediction": {
+                "type": "probability",
+                "format": "classification",
+                "num_channels": 2,
+                "spatial_shape": [
+                    "164",
+                    "164"
+                ],
+                "dtype": "float32",
+                "value_range": [
+                    0,
+                    1
+                ],
+                "is_patch_data": true,
+                "channel_def": {
+                    "0": "background",
+                    "1": "type of nucleus for each pixel"
+                }
+            }
+        }
+    }
+}

configs/multi_gpu_train.json

@@ -0,0 +1,36 @@
+{
+    "device": "$torch.device(f'cuda:{dist.get_rank()}')",
+    "network": {
+        "_target_": "torch.nn.parallel.DistributedDataParallel",
+        "module": "$@network_def.to(@device)",
+        "device_ids": [
+            "@device"
+        ]
+    },
+    "train#sampler": {
+        "_target_": "DistributedSampler",
+        "dataset": "@train#dataset",
+        "even_divisible": true,
+        "shuffle": true
+    },
+    "train#dataloader#sampler": "@train#sampler",
+    "train#dataloader#shuffle": false,
+    "train#trainer#train_handlers": "$@train#train_handlers[: -2 if dist.get_rank() > 0 else None]",
+    "validate#sampler": {
+        "_target_": "DistributedSampler",
+        "dataset": "@validate#dataset",
+        "even_divisible": false,
+        "shuffle": false
+    },
+    "validate#dataloader#sampler": "@validate#sampler",
+    "validate#evaluator#val_handlers": "$None if dist.get_rank() > 0 else @validate#handlers",
+    "training": [
+        "$import torch.distributed as dist",
+        "$dist.init_process_group(backend='nccl')",
+        "$torch.cuda.set_device(@device)",
+        "$monai.utils.set_determinism(seed=321)",
+        "$setattr(torch.backends.cudnn, 'benchmark', True)",
+        "$@train#trainer.run()",
+        "$dist.destroy_process_group()"
+    ]
+}

configs/train.json

@@ -0,0 +1,525 @@
+{
+    "imports": [
+        "$import glob",
+        "$import os",
+        "$import skimage"
+    ],
+    "bundle_root": "$os.getcwd()",
+    "ckpt_dir_stage0": "$os.path.join(@bundle_root, 'models', 'stage0')",
+    "ckpt_dir_stage1": "$os.path.join(@bundle_root, 'models')",
+    "ckpt_path_stage0": "$os.path.join(@ckpt_dir_stage0, 'model.pt')",
+    "output_dir": "$os.path.join(@bundle_root, 'eval')",
+    "dataset_dir": "/workspace/Data/Pathology/CoNSeP/Prepared/",
+    "train_images": "$list(sorted(glob.glob(@dataset_dir + '/Train/*image.npy')))",
+    "val_images": "$list(sorted(glob.glob(@dataset_dir + '/Test/*image.npy')))",
+    "train_inst_map": "$list(sorted(glob.glob(@dataset_dir + '/Train/*inst_map.npy')))",
+    "val_inst_map": "$list(sorted(glob.glob(@dataset_dir + '/Test/*inst_map.npy')))",
+    "train_type_map": "$list(sorted(glob.glob(@dataset_dir + '/Train/*type_map.npy')))",
+    "val_type_map": "$list(sorted(glob.glob(@dataset_dir + '/Test/*type_map.npy')))",
+    "device": "$torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')",
+    "stage": 0,
+    "epochs": 50,
+    "val_interval": 1,
+    "learning_rate": 0.0001,
+    "amp": true,
+    "hovernet_mode": "fast",
+    "patch_size": 256,
+    "out_size": 164,
+    "ckpt_dir": "$@ckpt_dir_stage0 if @stage == 0 else @ckpt_dir_stage1",
+    "network_def": {
+        "_target_": "HoVerNet",
+        "mode": "@hovernet_mode",
+        "in_channels": 3,
+        "out_classes": 5,
+        "adapt_standard_resnet": true,
+        "pretrained_url": "$None",
+        "freeze_encoder": true
+    },
+    "network": "$@network_def.to(@device)",
+    "loss": {
+        "_target_": "HoVerNetLoss",
+        "lambda_hv_mse": 1.0
+    },
+    "optimizer": {
+        "_target_": "torch.optim.Adam",
+        "params": "$filter(lambda p: p.requires_grad, @network.parameters())",
+        "lr": "@learning_rate",
+        "weight_decay": 1e-05
+    },
+    "lr_scheduler": {
+        "_target_": "torch.optim.lr_scheduler.StepLR",
+        "optimizer": "@optimizer",
+        "step_size": 25
+    },
+    "train": {
+        "preprocessing_transforms": [
+            {
+                "_target_": "LoadImaged",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ]
+            },
+            {
+                "_target_": "EnsureChannelFirstd",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "channel_dim": -1
+            },
+            {
+                "_target_": "Lambdad",
+                "keys": "label_inst",
+                "func": "$lambda x: skimage.measure.label(x)"
+            },
+            {
+                "_target_": "RandAffined",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "prob": 1.0,
+                "rotate_range": [
+                    "$np.pi"
+                ],
+                "scale_range": [
+                    [
+                        -0.2,
+                        0.2
+                    ],
+                    [
+                        -0.2,
+                        0.2
+                    ]
+                ],
+                "shear_range": [
+                    [
+                        -0.05,
+                        0.05
+                    ],
+                    [
+                        -0.05,
+                        0.05
+                    ]
+                ],
+                "translate_range": [
+                    [
+                        -6,
+                        6
+                    ],
+                    [
+                        -6,
+                        6
+                    ]
+                ],
+                "padding_mode": "zeros",
+                "mode": "nearest"
+            },
+            {
+                "_target_": "CenterSpatialCropd",
+                "keys": [
+                    "image"
+                ],
+                "roi_size": [
+                    "@patch_size",
+                    "@patch_size"
+                ]
+            },
+            {
+                "_target_": "RandFlipd",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "prob": 0.5,
+                "spatial_axis": 0
+            },
+            {
+                "_target_": "RandFlipd",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "prob": 0.5,
+                "spatial_axis": 1
+            },
+            {
+                "_target_": "OneOf",
+                "transforms": [
+                    {
+                        "_target_": "RandGaussianSmoothd",
+                        "keys": [
+                            "image"
+                        ],
+                        "sigma_x": [
+                            0.1,
+                            1.1
+                        ],
+                        "sigma_y": [
+                            0.1,
+                            1.1
+                        ],
+                        "prob": 1.0
+                    },
+                    {
+                        "_target_": "MedianSmoothd",
+                        "keys": [
+                            "image"
+                        ],
+                        "radius": 1
+                    },
+                    {
+                        "_target_": "RandGaussianNoised",
+                        "keys": [
+                            "image"
+                        ],
+                        "std": 0.05,
+                        "prob": 1.0
+                    }
+                ]
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": "image",
+                "dtype": "$np.uint8"
+            },
+            {
+                "_target_": "TorchVisiond",
+                "keys": "image",
+                "name": "ColorJitter",
+                "brightness": [
+                    0.9,
+                    1.0
+                ],
+                "contrast": [
+                    0.95,
+                    1.1
+                ],
+                "saturation": [
+                    0.8,
+                    1.2
+                ],
+                "hue": [
+                    -0.04,
+                    0.04
+                ]
+            },
+            {
+                "_target_": "AsDiscreted",
+                "keys": "label_type",
+                "to_onehot": 5
+            },
+            {
+                "_target_": "ScaleIntensityRanged",
+                "keys": "image",
+                "a_min": 0.0,
+                "a_max": 255.0,
+                "b_min": 0.0,
+                "b_max": 1.0,
+                "clip": true
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": "label_inst",
+                "dtype": "$torch.int"
+            },
+            {
+                "_target_": "ComputeHoVerMapsd",
+                "keys": "label_inst"
+            },
+            {
+                "_target_": "Lambdad",
+                "keys": "label_inst",
+                "func": "$lambda x: x > 0",
+                "overwrite": "label"
+            },
+            {
+                "_target_": "CenterSpatialCropd",
+                "keys": [
+                    "label",
+                    "hover_label_inst",
+                    "label_inst",
+                    "label_type"
+                ],
+                "roi_size": [
+                    "@out_size",
+                    "@out_size"
+                ]
+            },
+            {
+                "_target_": "AsDiscreted",
+                "keys": "label",
+                "to_onehot": 2
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "dtype": "$torch.float32"
+            }
+        ],
+        "preprocessing": {
+            "_target_": "Compose",
+            "transforms": "$@train#preprocessing_transforms"
+        },
+        "dataset": {
+            "_target_": "Dataset",
+            "data": "$[{'image': i, 'label_inst': j, 'label_type': k} for i, j, k in zip(@train_images, @train_inst_map, @train_type_map)]",
+            "transform": "@train#preprocessing"
+        },
+        "dataloader": {
+            "_target_": "DataLoader",
+            "dataset": "@train#dataset",
+            "batch_size": 16,
+            "shuffle": true,
+            "num_workers": 4
+        },
+        "inferer": {
+            "_target_": "SimpleInferer"
+        },
+        "postprocessing_np": {
+            "_target_": "Compose",
+            "transforms": [
+                {
+                    "_target_": "Activationsd",
+                    "keys": "nucleus_prediction",
+                    "softmax": true
+                },
+                {
+                    "_target_": "AsDiscreted",
+                    "keys": "nucleus_prediction",
+                    "argmax": true
+                }
+            ]
+        },
+        "postprocessing": {
+            "_target_": "Lambdad",
+            "keys": "pred",
+            "func": "$@train#postprocessing_np"
+        },
+        "handlers": [
+            {
+                "_target_": "LrScheduleHandler",
+                "lr_scheduler": "@lr_scheduler",
+                "print_lr": true
+            },
+            {
+                "_target_": "ValidationHandler",
+                "validator": "@validate#evaluator",
+                "epoch_level": true,
+                "interval": "@val_interval"
+            },
+            {
+                "_target_": "CheckpointSaver",
+                "save_dir": "@ckpt_dir",
+                "save_dict": {
+                    "model": "@network"
+                },
+                "save_interval": 10,
+                "epoch_level": true,
+                "save_final": true,
+                "final_filename": "model.pt"
+            },
+            {
+                "_target_": "StatsHandler",
+                "tag_name": "train_loss",
+                "output_transform": "$monai.handlers.from_engine(['loss'], first=True)"
+            },
+            {
+                "_target_": "TensorBoardStatsHandler",
+                "log_dir": "@output_dir",
+                "tag_name": "train_loss",
+                "output_transform": "$monai.handlers.from_engine(['loss'], first=True)"
+            }
+        ],
+        "extra_handlers": [
+            {
+                "_target_": "CheckpointLoader",
+                "load_path": "$os.path.join(@ckpt_dir_stage0, 'model.pt')",
+                "load_dict": {
+                    "model": "@network"
+                }
+            }
+        ],
+        "train_handlers": "$@train#extra_handlers + @train#handlers if @stage==1 else @train#handlers",
+        "key_metric": {
+            "train_mean_dice": {
+                "_target_": "MeanDice",
+                "include_background": false,
+                "output_transform": "$monai.apps.pathology.handlers.utils.from_engine_hovernet(keys=['pred', 'label'], nested_key='nucleus_prediction')"
+            }
+        },
+        "trainer": {
+            "_target_": "SupervisedTrainer",
+            "max_epochs": "@epochs",
+            "device": "@device",
+            "train_data_loader": "@train#dataloader",
+            "prepare_batch": "$monai.apps.pathology.engines.utils.PrepareBatchHoVerNet(extra_keys=['label_type', 'hover_label_inst'])",
+            "network": "@network",
+            "loss_function": "@loss",
+            "optimizer": "@optimizer",
+            "inferer": "@train#inferer",
+            "postprocessing": "@train#postprocessing",
+            "key_train_metric": "@train#key_metric",
+            "train_handlers": "@train#train_handlers",
+            "amp": "@amp"
+        }
+    },
+    "validate": {
+        "preprocessing_transforms": [
+            {
+                "_target_": "LoadImaged",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ]
+            },
+            {
+                "_target_": "EnsureChannelFirstd",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "channel_dim": -1
+            },
+            {
+                "_target_": "Lambdad",
+                "keys": "label_inst",
+                "func": "$lambda x: skimage.measure.label(x)"
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": [
+                    "image",
+                    "label_inst"
+                ],
+                "dtype": "$torch.int"
+            },
+            {
+                "_target_": "CenterSpatialCropd",
+                "keys": [
+                    "image"
+                ],
+                "roi_size": [
+                    "@patch_size",
+                    "@patch_size"
+                ]
+            },
+            {
+                "_target_": "ScaleIntensityRanged",
+                "keys": "image",
+                "a_min": 0.0,
+                "a_max": 255.0,
+                "b_min": 0.0,
+                "b_max": 1.0,
+                "clip": true
+            },
+            {
+                "_target_": "ComputeHoVerMapsd",
+                "keys": "label_inst"
+            },
+            {
+                "_target_": "Lambdad",
+                "keys": "label_inst",
+                "func": "$lambda x: x > 0",
+                "overwrite": "label"
+            },
+            {
+                "_target_": "CenterSpatialCropd",
+                "keys": [
+                    "label",
+                    "hover_label_inst",
+                    "label_inst",
+                    "label_type"
+                ],
+                "roi_size": [
+                    "@out_size",
+                    "@out_size"
+                ]
+            },
+            {
+                "_target_": "CastToTyped",
+                "keys": [
+                    "image",
+                    "label_inst",
+                    "label_type"
+                ],
+                "dtype": "$torch.float32"
+            }
+        ],
+        "preprocessing": {
+            "_target_": "Compose",
+            "transforms": "$@validate#preprocessing_transforms"
+        },
+        "dataset": {
+            "_target_": "Dataset",
+            "data": "$[{'image': i, 'label_inst': j, 'label_type': k} for i, j, k in zip(@val_images, @val_inst_map, @val_type_map)]",
+            "transform": "@validate#preprocessing"
+        },
+        "dataloader": {
+            "_target_": "DataLoader",
+            "dataset": "@validate#dataset",
+            "batch_size": 16,
+            "shuffle": false,
+            "num_workers": 4
+        },
+        "inferer": {
+            "_target_": "SimpleInferer"
+        },
+        "postprocessing": "$@train#postprocessing",
+        "handlers": [
+            {
+                "_target_": "StatsHandler",
+                "iteration_log": false
+            },
+            {
+                "_target_": "TensorBoardStatsHandler",
+                "log_dir": "@output_dir",
+                "iteration_log": false
+            },
+            {
+                "_target_": "CheckpointSaver",
+                "save_dir": "@ckpt_dir",
+                "save_dict": {
+                    "model": "@network"
+                },
+                "save_key_metric": true
+            }
+        ],
+        "key_metric": {
+            "val_mean_dice": {
+                "_target_": "MeanDice",
+                "include_background": false,
+                "output_transform": "$monai.apps.pathology.handlers.utils.from_engine_hovernet(keys=['pred', 'label'], nested_key='nucleus_prediction')"
+            }
+        },
+        "evaluator": {
+            "_target_": "SupervisedEvaluator",
+            "device": "@device",
+            "val_data_loader": "@validate#dataloader",
+            "prepare_batch": "$monai.apps.pathology.engines.utils.PrepareBatchHoVerNet(extra_keys=['label_type', 'hover_label_inst'])",
+            "network": "@network",
+            "inferer": "@validate#inferer",
+            "postprocessing": "@validate#postprocessing",
+            "key_val_metric": "@validate#key_metric",
+            "val_handlers": "@validate#handlers",
+            "amp": "@amp"
+        }
+    },
+    "training": [
+        "$monai.utils.set_determinism(seed=321)",
+        "$setattr(torch.backends.cudnn, 'benchmark', True)",
+        "$@train#trainer.run()"
+    ]
+}

docs/README.md

@@ -0,0 +1,147 @@
+# Model Overview
+
+A pre-trained model for simultaneous segmentation and classification of nuclei within multi-tissue histology images based on CoNSeP data. The details of the model can be found in [1].
+
+## Workflow
+
+The model is trained to simultaneous segment and classify nuclei. Training is done via a two-stage approach. First initialized the model with pre-trained weights on the [ImageNet dataset](https://ieeexplore.ieee.org/document/5206848), trained only the decoders for the first 50 epochs, and then fine-tuned all layers for another 50 epochs. There are two training modes in total. If "original" mode is specified, it uses [270, 270] and [80, 80] for `patch_size` and `out_size` respectively. If "fast" mode is specified, it uses [256, 256] and [164, 164] for `patch_size` and `out_size` respectively. The results we show below are based on the "fast" model.
+
+- We train the first stage with pre-trained weights from some internal data.
+
+- The original author's repo also has pre-trained weights which is for non-commercial use. Each user is responsible for checking the content of models/datasets and the applicable licenses and determining if suitable for the intended use. The license for the pre-trained model is different than MONAI license. Please check the source where these weights are obtained from: <https://github.com/vqdang/hover_net#data-format>
+
+`PRETRAIN_MODEL_URL` is "https://drive.google.com/u/1/uc?id=1KntZge40tAHgyXmHYVqZZ5d2p_4Qr2l5&export=download" which can be used in bash code below.
+
+![Model workflow](https://ars.els-cdn.com/content/image/1-s2.0-S1361841519301045-fx1_lrg.jpg)
+
+## Data
+
+The training data is from <https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/>.
+
+- Target: segment instance-level nuclei and classify the nuclei type
+- Task: Segmentation and classification
+- Modality: RGB images
+- Size: 41 image tiles (2009 patches)
+
+The provided labelled data was partitioned, based on the original split, into training (27 tiles) and testing (14 tiles) datasets.
+
+After download the datasets, please run `scripts/prepare_patches.py` to prepare patches from tiles. Prepared patches are saved in `your-concep-dataset-path`/Prepared. The implementation is referring to <https://github.com/vqdang/hover_net/blob/master/extract_patches.py>. The command is like:
+
+```
+python scripts/prepare_patches.py -root your-concep-dataset-path
+```
+
+## Training configuration
+
+This model utilized a two-stage approach. The training was performed with the following:
+
+- GPU: At least 24GB of GPU memory.
+- Actual Model Input: 256 x 256
+- AMP: True
+- Optimizer: Adam
+- Learning Rate: 1e-4
+- Loss: HoVerNetLoss
+
+## Input
+
+Input: RGB images
+
+## Output
+
+Output: a dictionary with the following keys:
+
+1. nucleus_prediction: predict whether or not a pixel belongs to the nuclei or background
+2. horizontal_vertical: predict the horizontal and vertical distances of nuclear pixels to their centres of mass
+3. type_prediction: predict the type of nucleus for each pixel
+
+## Model Performance
+
+The achieved metrics on the validation data are:
+
+Fast mode:
+- Binary Dice: 0.8293
+- PQ: 0.4936
+- F1d: 0.7480
+
+#### Training Loss and Dice
+
+stage1:
+![A graph showing the training loss and the mean dice over 50 epochs in stage1](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_train_stage1_fast.png)
+
+stage2:
+![A graph showing the training loss and the mean dice over 50 epochs in stage2](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_train_stage2_fast.png)
+
+#### Validation Dice
+
+stage1:
+
+![A graph showing the validation mean dice over 50 epochs in stage1](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_val_stage1_fast.png)
+
+stage2:
+
+![A graph showing the validation mean dice over 50 epochs in stage2](https://developer.download.nvidia.com/assets/Clara/Images/monai_pathology_nuclei_seg_cls_val_stage2_fast.png)
+
+## commands example
+
+Execute training:
+
+- Run first stage
+
+```
+python -m monai.bundle run training --meta_file configs/metadata.json --config_file configs/train.json --logging_file configs/logging.conf --network_def#pretrained_url `PRETRAIN_MODEL_URL` --stage 0
+```
+
+- Run second stage
+
+```
+python -m monai.bundle run training --meta_file configs/metadata.json --config_file configs/train.json --logging_file configs/logging.conf --network_def#freeze_encoder false --network_def#pretrained_url None --stage 1
+```
+
+Override the `train` config to execute multi-GPU training:
+
+- Run first stage
+
+```
+torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run training --meta_file configs/metadata.json --config_file "['configs/train.json','configs/multi_gpu_train.json']" --logging_file configs/logging.conf --train#dataloader#batch_size 8 --network_def#freeze_encoder true --network_def#pretrained_url `PRETRAIN_MODEL_URL` --stage 0
+```
+
+- Run second stage
+
+```
+torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run training --meta_file configs/metadata.json --config_file "['configs/train.json','configs/multi_gpu_train.json']" --logging_file configs/logging.conf --train#dataloader#batch_size 4 --network_def#freeze_encoder false --network_def#pretrained_url None --stage 1
+```
+
+Override the `train` config to execute evaluation with the trained model:
+
+```
+python -m monai.bundle run evaluating --meta_file configs/metadata.json --config_file "['configs/train.json','configs/evaluate.json']" --logging_file configs/logging.conf
+```
+
+### Execute inference
+
+```
+python -m monai.bundle run evaluating --meta_file configs/metadata.json --config_file configs/inference.json --logging_file configs/logging.conf
+```
+
+# Disclaimer
+
+This is an example, not to be used for diagnostic purposes.
+
+# References
+
+[1] Simon Graham, Quoc Dang Vu, Shan E Ahmed Raza, Ayesha Azam, Yee Wah Tsang, Jin Tae Kwak, Nasir Rajpoot, Hover-Net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images, Medical Image Analysis, 2019 https://doi.org/10.1016/j.media.2019.101563
+
+# License
+Copyright (c) MONAI Consortium
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

docs/data_license.txt

@@ -0,0 +1,6 @@
+Third Party Licenses
+-----------------------------------------------------------------------
+
+/*********************************************************************/
+i. CoNSeP dataset
+   https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/

models/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3aa9651e11eca4c17d89fe59b46b8a51c7899decb03df538f5092ee9e55967ef
+size 151214500

models/stage0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a394593bc2ed0ef53188ffc96174ae492b59448a37b4b4a161214a30b173785
+size 151220093

scripts/prepare_patches.py

@@ -0,0 +1,232 @@
+import glob
+import math
+import os
+import pathlib
+import shutil
+from argparse import ArgumentParser
+
+import numpy as np
+import scipy.io as sio
+import tqdm
+from PIL import Image
+
+
+def load_img(path):
+    return np.array(Image.open(path).convert("RGB"))
+
+
+def load_ann(path):
+    """
+    This function is specific to CoNSeP dataset.
+    If using other datasets, the code below may need to be modified.
+    """
+    # assumes that ann is HxW
+    ann_inst = sio.loadmat(path)["inst_map"]
+    ann_type = sio.loadmat(path)["type_map"]
+
+    # merge classes for CoNSeP (utilise 3 nuclei classes and background keep the same with paper)
+    ann_type[(ann_type == 3) | (ann_type == 4)] = 3
+    ann_type[(ann_type == 5) | (ann_type == 6) | (ann_type == 7)] = 4
+
+    ann = np.dstack([ann_inst, ann_type])
+    ann = ann.astype("int32")
+
+    return ann
+
+
+class PatchExtractor:
+    """Extractor to generate patches with or without padding.
+    Turn on debug mode to see how it is done.
+
+    Args:
+        x         : input image, should be of shape HWC
+        patch_size  : a tuple of (h, w)
+        step_size : a tuple of (h, w)
+    Return:
+        a list of sub patches, each patch has dtype same as x
+
+    Examples:
+        >>> xtractor = PatchExtractor((450, 450), (120, 120))
+        >>> img = np.full([1200, 1200, 3], 255, np.uint8)
+        >>> patches = xtractor.extract(img, 'mirror')
+
+    """
+
+    def __init__(self, patch_size, step_size):
+        self.patch_type = "mirror"
+        self.patch_size = patch_size
+        self.step_size = step_size
+
+    def __get_patch(self, x, ptx):
+        pty = (ptx[0] + self.patch_size[0], ptx[1] + self.patch_size[1])
+        win = x[ptx[0] : pty[0], ptx[1] : pty[1]]
+        assert (
+            win.shape[0] == self.patch_size[0] and win.shape[1] == self.patch_size[1]
+        ), "[BUG] Incorrect Patch Size {0}".format(win.shape)
+        return win
+
+    def __extract_valid(self, x):
+        """Extracted patches without padding, only work in case patch_size > step_size.
+
+        Note: to deal with the remaining portions which are at the boundary a.k.a
+        those which do not fit when slide left->right, top->bottom), we flip
+        the sliding direction then extract 1 patch starting from right / bottom edge.
+        There will be 1 additional patch extracted at the bottom-right corner.
+
+        Args:
+            x         : input image, should be of shape HWC
+            patch_size  : a tuple of (h, w)
+            step_size : a tuple of (h, w)
+        Return:
+            a list of sub patches, each patch is same dtype as x
+
+        """
+        im_h = x.shape[0]
+        im_w = x.shape[1]
+
+        def extract_infos(length, patch_size, step_size):
+            flag = (length - patch_size) % step_size != 0
+            last_step = math.floor((length - patch_size) / step_size)
+            last_step = (last_step + 1) * step_size
+            return flag, last_step
+
+        h_flag, h_last = extract_infos(im_h, self.patch_size[0], self.step_size[0])
+        w_flag, w_last = extract_infos(im_w, self.patch_size[1], self.step_size[1])
+
+        sub_patches = []
+        # Deal with valid block
+        for row in range(0, h_last, self.step_size[0]):
+            for col in range(0, w_last, self.step_size[1]):
+                win = self.__get_patch(x, (row, col))
+                sub_patches.append(win)
+        # Deal with edge case
+        if h_flag:
+            row = im_h - self.patch_size[0]
+            for col in range(0, w_last, self.step_size[1]):
+                win = self.__get_patch(x, (row, col))
+                sub_patches.append(win)
+        if w_flag:
+            col = im_w - self.patch_size[1]
+            for row in range(0, h_last, self.step_size[0]):
+                win = self.__get_patch(x, (row, col))
+                sub_patches.append(win)
+        if h_flag and w_flag:
+            ptx = (im_h - self.patch_size[0], im_w - self.patch_size[1])
+            win = self.__get_patch(x, ptx)
+            sub_patches.append(win)
+        return sub_patches
+
+    def __extract_mirror(self, x):
+        """Extracted patches with mirror padding the boundary such that the
+        central region of each patch is always within the orginal (non-padded)
+        image while all patches' central region cover the whole orginal image.
+
+        Args:
+            x         : input image, should be of shape HWC
+            patch_size  : a tuple of (h, w)
+            step_size : a tuple of (h, w)
+        Return:
+            a list of sub patches, each patch is same dtype as x
+
+        """
+        diff_h = self.patch_size[0] - self.step_size[0]
+        padt = diff_h // 2
+        padb = diff_h - padt
+
+        diff_w = self.patch_size[1] - self.step_size[1]
+        padl = diff_w // 2
+        padr = diff_w - padl
+
+        pad_type = "reflect"
+        x = np.lib.pad(x, ((padt, padb), (padl, padr), (0, 0)), pad_type)
+        sub_patches = self.__extract_valid(x)
+        return sub_patches
+
+    def extract(self, x, patch_type):
+        patch_type = patch_type.lower()
+        self.patch_type = patch_type
+        if patch_type == "valid":
+            return self.__extract_valid(x)
+        elif patch_type == "mirror":
+            return self.__extract_mirror(x)
+        else:
+            raise ValueError(f"Unknown Patch Type {patch_type}")
+
+
+def main(cfg):
+    xtractor = PatchExtractor(cfg["patch_size"], cfg["step_size"])
+    for phase in cfg["phase"]:
+        img_dir = os.path.join(cfg["root"], f"{phase}/Images")
+        ann_dir = os.path.join(cfg["root"], f"{phase}/Labels")
+
+        file_list = glob.glob(os.path.join(ann_dir, f"*{cfg['label_suffix']}"))
+        file_list.sort()  # ensure same ordering across platform
+
+        out_dir = f"{cfg['root']}/Prepared/{phase}"
+        if os.path.isdir(out_dir):
+            shutil.rmtree(out_dir)
+        os.makedirs(out_dir)
+
+        pbar_format = "Process File: |{bar}| {n_fmt}/{total_fmt}[{elapsed}<{remaining},{rate_fmt}]"
+        pbarx = tqdm.tqdm(total=len(file_list), bar_format=pbar_format, ascii=True, position=0)
+
+        for file_path in file_list:
+            base_name = pathlib.Path(file_path).stem
+
+            img = load_img(f"{img_dir}/{base_name}.{cfg['image_suffix']}")
+            ann = load_ann(f"{ann_dir}/{base_name}.{cfg['label_suffix']}")
+
+            # *
+            img = np.concatenate([img, ann], axis=-1)
+            sub_patches = xtractor.extract(img, cfg["extract_type"])
+
+            pbar_format = "Extracting  : |{bar}| {n_fmt}/{total_fmt}[{elapsed}<{remaining},{rate_fmt}]"
+            pbar = tqdm.tqdm(total=len(sub_patches), leave=False, bar_format=pbar_format, ascii=True, position=1)
+
+            for idx, patch in enumerate(sub_patches):
+                image_patch = patch[..., :3]
+                inst_map_patch = patch[..., 3:4]
+                type_map_patch = patch[..., 4:5]
+                np.save("{0}/{1}_{2:03d}_image.npy".format(out_dir, base_name, idx), image_patch)
+                np.save("{0}/{1}_{2:03d}_inst_map.npy".format(out_dir, base_name, idx), inst_map_patch)
+                np.save("{0}/{1}_{2:03d}_type_map.npy".format(out_dir, base_name, idx), type_map_patch)
+                pbar.update()
+            pbar.close()
+            # *
+
+            pbarx.update()
+        pbarx.close()
+
+
+def parse_arguments():
+    parser = ArgumentParser(description="Extract patches from the original images")
+
+    parser.add_argument(
+        "--root",
+        type=str,
+        default="/workspace/Data/Pathology/CoNSeP",
+        help="root path to image folder containing training/test",
+    )
+    parser.add_argument(
+        "--phase",
+        nargs="+",
+        type=str,
+        default=["Train", "Test"],
+        dest="phase",
+        help="Phases of data need to be extracted",
+    )
+    parser.add_argument("--type", type=str, default="mirror", dest="extract_type", help="Choose 'mirror' or 'valid'")
+    parser.add_argument("--is", type=str, default="png", dest="image_suffix", help="image file name suffix")
+    parser.add_argument("--ls", type=str, default="mat", dest="label_suffix", help="label file name suffix")
+    parser.add_argument("--ps", nargs="+", type=int, default=[540, 540], dest="patch_size", help="patch size")
+    parser.add_argument("--ss", nargs="+", type=int, default=[164, 164], dest="step_size", help="patch size")
+    args = parser.parse_args()
+    config_dict = vars(args)
+
+    return config_dict
+
+
+if __name__ == "__main__":
+    cfg = parse_arguments()
+
+    main(cfg)