From a6960f8799bd3837d68b3216b2157821dda8894e Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Thu, 16 Feb 2023 13:10:30 +0000
Subject: [PATCH 01/14] :sparkles: NEW: Add Nucleus Detection Engine - Add
 Nucleus Detection Engine - Fix documentation in patch_predictor.py

---
 tiatoolbox/models/engine/__init__.py          |   1 +
 tiatoolbox/models/engine/nucleus_detection.py | 136 ++++++++++++++++++
 tiatoolbox/models/engine/patch_predictor.py   |  38 ++---
 3 files changed, 157 insertions(+), 18 deletions(-)
 create mode 100644 tiatoolbox/models/engine/nucleus_detection.py

diff --git a/tiatoolbox/models/engine/__init__.py b/tiatoolbox/models/engine/__init__.py
index 2cba98a32..7574d6e7e 100644
--- a/tiatoolbox/models/engine/__init__.py
+++ b/tiatoolbox/models/engine/__init__.py
@@ -1,5 +1,6 @@
 """Engines to run models implemented in tiatoolbox."""
 from tiatoolbox.models.engine import (
+    nucleus_detection,
     nucleus_instance_segmentor,
     patch_predictor,
     semantic_segmentor,
diff --git a/tiatoolbox/models/engine/nucleus_detection.py b/tiatoolbox/models/engine/nucleus_detection.py
new file mode 100644
index 000000000..bca6aa20a
--- /dev/null
+++ b/tiatoolbox/models/engine/nucleus_detection.py
@@ -0,0 +1,136 @@
+"""This module implements nucleus detection engine."""
+
+
+from tiatoolbox.models.architecture import get_pretrained_model
+from tiatoolbox.models.engine.patch_predictor import PatchPredictor
+
+
+class NucleusDetector(PatchPredictor):
+    r"""Nucleus detection engine.
+
+    The models provided by tiatoolbox should give the following results:
+
+    .. list-table:: Nucleus detection performance on the (add models list here)
+       :widths: 15 15
+       :header-rows: 1
+
+    Args:
+        model (nn.Module):
+            Use externally defined PyTorch model for prediction with.
+            weights already loaded. Default is `None`. If provided,
+            `pretrained_model` argument is ignored.
+        pretrained_model (str):
+            Name of the existing models support by tiatoolbox for
+            processing the data. For a full list of pretrained models,
+            refer to the `docs
+            <https://tia-toolbox.readthedocs.io/en/latest/pretrained.html>`_
+            By default, the corresponding pretrained weights will also
+            be downloaded. However, you can override with your own set
+            of weights via the `pretrained_weights` argument. Argument
+            is case-insensitive.
+        pretrained_weights (str):
+            Path to the weight of the corresponding `pretrained_model`.
+
+          >>> predictor = NucleusDetector(
+          ...    pretrained_model="resnet18-kather100k",
+          ...    pretrained_weights="resnet18_local_weight")
+
+        batch_size (int):
+            Number of images fed into the model each time.
+        num_loader_workers (int):
+            Number of workers to load the data. Take note that they will
+            also perform preprocessing.
+        verbose (bool):
+            Whether to output logging information.
+
+    Attributes:
+        imgs (:obj:`str` or :obj:`pathlib.Path` or :obj:`numpy.ndarray`):
+            A HWC image or a path to WSI.
+        mode (str):
+            Type of input to process. Choose from either `patch`, `tile`
+            or `wsi`.
+        model (nn.Module):
+            Defined PyTorch model.
+        pretrained_model (str):
+            Name of the existing models support by tiatoolbox for
+            processing the data. For a full list of pretrained models,
+            refer to the `docs
+            <https://tia-toolbox.readthedocs.io/en/latest/pretrained.html>`_
+            By default, the corresponding pretrained weights will also
+            be downloaded. However, you can override with your own set
+            of weights via the `pretrained_weights` argument. Argument
+            is case insensitive.
+        batch_size (int):
+            Number of images fed into the model each time.
+        num_loader_workers (int):
+            Number of workers used in torch.utils.data.DataLoader.
+        verbose (bool):
+            Whether to output logging information.
+
+    Examples:
+        >>> # list of 2 image patches as input
+        >>> data = [img1, img2]
+        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
+        >>> output = predictor.predict(data, mode='patch')
+
+        >>> # array of list of 2 image patches as input
+        >>> data = np.array([img1, img2])
+        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
+        >>> output = predictor.predict(data, mode='patch')
+
+        >>> # list of 2 image patch files as input
+        >>> data = ['path/img.png', 'path/img.png']
+        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
+        >>> output = predictor.predict(data, mode='patch')
+
+        >>> # list of 2 image tile files as input
+        >>> tile_file = ['path/tile1.png', 'path/tile2.png']
+        >>> predictor = PatchPredictor(pretraind_model="resnet18-kather100k")
+        >>> output = predictor.predict(tile_file, mode='tile')
+
+        >>> # list of 2 wsi files as input
+        >>> wsi_file = ['path/wsi1.svs', 'path/wsi2.svs']
+        >>> predictor = PatchPredictor(pretraind_model="resnet18-kather100k")
+        >>> output = predictor.predict(wsi_file, mode='wsi')
+
+    References:
+        [1] Kather, Jakob Nikolas, et al. "Predicting survival from colorectal cancer
+        histology slides using deep learning: A retrospective multicenter study."
+        PLoS medicine 16.1 (2019): e1002730.
+
+        [2] Veeling, Bastiaan S., et al. "Rotation equivariant CNNs for digital
+        pathology." International Conference on Medical image computing and
+        computer-assisted intervention. Springer, Cham, 2018.
+
+    """  # noqa: W605
+
+    def __init__(
+        self,
+        batch_size=8,
+        num_loader_workers=0,
+        model=None,
+        pretrained_model=None,
+        pretrained_weights=None,
+        verbose=True,
+    ):
+        super().__init__()
+
+        self.imgs = None
+        self.mode = None
+
+        if model is None and pretrained_model is None:
+            raise ValueError("Must provide either `model` or `pretrained_model`.")
+
+        if model is not None:
+            self.model = model
+            ioconfig = None  # retrieve iostate from provided model ?
+        else:
+            model, ioconfig = get_pretrained_model(pretrained_model, pretrained_weights)
+
+        self.ioconfig = ioconfig  # for storing original
+        self._ioconfig = None  # for storing runtime
+        self.model = model  # for runtime, such as after wrapping with nn.DataParallel
+        self.pretrained_model = pretrained_model
+        self.batch_size = batch_size
+        self.num_loader_worker = num_loader_workers
+        self.verbose = verbose
diff --git a/tiatoolbox/models/engine/patch_predictor.py b/tiatoolbox/models/engine/patch_predictor.py
index f8bccdf84..d9ef18312 100644
--- a/tiatoolbox/models/engine/patch_predictor.py
+++ b/tiatoolbox/models/engine/patch_predictor.py
@@ -246,7 +246,7 @@ def __init__(
         self.model = model  # for runtime, such as after wrapping with nn.DataParallel
         self.pretrained_model = pretrained_model
         self.batch_size = batch_size
-        self.num_loader_worker = num_loader_workers
+        self.num_loader_workers = num_loader_workers
         self.verbose = verbose
 
     @staticmethod
@@ -398,7 +398,7 @@ def _predict_engine(
         # preprocessing must be defined with the dataset
         dataloader = torch.utils.data.DataLoader(
             dataset,
-            num_workers=self.num_loader_worker,
+            num_workers=self.num_loader_workers,
             batch_size=self.batch_size,
             drop_last=False,
             shuffle=False,
@@ -463,22 +463,24 @@ def _update_ioconfig(
 
         Args:
             ioconfig (IOPatchPredictorConfig):
-        patch_input_shape (tuple):
-            Size of patches input to the model. Patches are at
-            requested read resolution, not with respect to level 0,
-            and must be positive.
-        stride_shape (tuple):
-            Stride using during tile and WSI processing. Stride is
-            at requested read resolution, not with respect to
-            level 0, and must be positive. If not provided,
-            `stride_shape=patch_input_shape`.
-        resolution (float):
-            Resolution used for reading the image. Please see
-            :obj:`WSIReader` for details.
-        units (str):
-            Units of resolution used for reading the image. Choose
-            from either `level`, `power` or `mpp`. Please see
-            :obj:`WSIReader` for details.
+                Object defines information about input and output placement
+                of patches for patch prediction.
+            patch_input_shape (tuple):
+                Size of patches input to the model. Patches are at
+                requested read resolution, not with respect to level 0,
+                and must be positive.
+            stride_shape (tuple):
+                Stride using during tile and WSI processing. Stride is
+                at requested read resolution, not with respect to
+                level 0, and must be positive. If not provided,
+                `stride_shape=patch_input_shape`.
+            resolution (float):
+                Resolution used for reading the image. Please see
+                :obj:`WSIReader` for details.
+            units (str):
+                Units of resolution used for reading the image. Choose
+                from either `level`, `power` or `mpp`. Please see
+                :obj:`WSIReader` for details.
 
         Returns:
             Updated Patch Predictor IO configuration.

From 6c4750666ae38d7252ffb35725c20ca30d0387f5 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Thu, 16 Feb 2023 13:25:05 +0000
Subject: [PATCH 02/14] :art: DEV: Improve structure of init - Improve
 structure of init

---
 tiatoolbox/models/engine/nucleus_detection.py | 30 +++++--------------
 tiatoolbox/models/engine/patch_predictor.py   |  2 +-
 2 files changed, 9 insertions(+), 23 deletions(-)

diff --git a/tiatoolbox/models/engine/nucleus_detection.py b/tiatoolbox/models/engine/nucleus_detection.py
index bca6aa20a..b31e11d7d 100644
--- a/tiatoolbox/models/engine/nucleus_detection.py
+++ b/tiatoolbox/models/engine/nucleus_detection.py
@@ -1,7 +1,6 @@
 """This module implements nucleus detection engine."""
 
 
-from tiatoolbox.models.architecture import get_pretrained_model
 from tiatoolbox.models.engine.patch_predictor import PatchPredictor
 
 
@@ -113,24 +112,11 @@ def __init__(
         pretrained_weights=None,
         verbose=True,
     ):
-        super().__init__()
-
-        self.imgs = None
-        self.mode = None
-
-        if model is None and pretrained_model is None:
-            raise ValueError("Must provide either `model` or `pretrained_model`.")
-
-        if model is not None:
-            self.model = model
-            ioconfig = None  # retrieve iostate from provided model ?
-        else:
-            model, ioconfig = get_pretrained_model(pretrained_model, pretrained_weights)
-
-        self.ioconfig = ioconfig  # for storing original
-        self._ioconfig = None  # for storing runtime
-        self.model = model  # for runtime, such as after wrapping with nn.DataParallel
-        self.pretrained_model = pretrained_model
-        self.batch_size = batch_size
-        self.num_loader_worker = num_loader_workers
-        self.verbose = verbose
+        super().__init__(
+            batch_size=batch_size,
+            num_loader_workers=num_loader_workers,
+            model=model,
+            pretrained_model=pretrained_model,
+            pretrained_weights=pretrained_weights,
+            verbose=verbose,
+        )
diff --git a/tiatoolbox/models/engine/patch_predictor.py b/tiatoolbox/models/engine/patch_predictor.py
index d9ef18312..4215c9794 100644
--- a/tiatoolbox/models/engine/patch_predictor.py
+++ b/tiatoolbox/models/engine/patch_predictor.py
@@ -158,7 +158,7 @@ class PatchPredictor:
             Whether to output logging information.
 
     Attributes:
-        img (:obj:`str` or :obj:`pathlib.Path` or :obj:`numpy.ndarray`):
+        imgs (:obj:`str` or :obj:`pathlib.Path` or :obj:`numpy.ndarray`):
             A HWC image or a path to WSI.
         mode (str):
             Type of input to process. Choose from either `patch`, `tile`

From 584e43a12bd424501e9909703166e91cf0eea13b Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Mon, 20 Feb 2023 16:57:17 +0000
Subject: [PATCH 03/14] :memo: DOC: Add documentation for mapde and sccnn - Add
 documentation for mapde and sccnn - Add a test for nucleus detection - Fix
 input in patch predictor test

---
 docs/pretrained.rst                           | 81 +++++++++++++++++++
 tests/models/test_nucleus_detection_engine.py |  8 ++
 tests/models/test_patch_predictor.py          |  4 +-
 tiatoolbox/data/pretrained_model.yaml         | 22 +++++
 4 files changed, 113 insertions(+), 2 deletions(-)
 create mode 100644 tests/models/test_nucleus_detection_engine.py

diff --git a/docs/pretrained.rst b/docs/pretrained.rst
index ec516fbe0..eb959b9b5 100644
--- a/docs/pretrained.rst
+++ b/docs/pretrained.rst
@@ -275,6 +275,87 @@ input output configuration:
 
     - hovernet_original_kumar
 
+Nucleus Detection
+^^^^^^^^^^^^^^^^^
+
+CRCHisto Dataset
+--------------
+
+We provide the following models trained using the `CoNSeP dataset <https://warwick.ac.uk/fac/cross_fac/tia/data/crchistolabelednucleihe//>`_, which uses the following
+input output configuration:
+
+.. collapse:: Input Output Configuration Details
+
+   .. code-block:: python
+
+        from tiatoolbox.models import IOPatchPredictorConfig
+        ioconfig = IOPatchPredictorConfig(
+            patch_input_shape=(31, 31),
+            stride_shape=(8, 8),
+            input_resolutions=[{"resolution": 0.25, "units": "mpp"}]
+        )
+
+
+.. collapse:: Model names
+
+    - sccnn-crchisto
+
+.. collapse:: Input Output Configuration Details
+
+   .. code-block:: python
+
+        from tiatoolbox.models import IOPatchPredictorConfig
+        ioconfig = IOPatchPredictorConfig(
+            patch_input_shape=(252, 252),
+            stride_shape=(150, 150),
+            input_resolutions=[{"resolution": 0.25, "units": "mpp"}]
+        )
+
+
+.. collapse:: Model names
+
+    - mapde-crchisto
+
+
+CoNIC Dataset
+--------------
+
+We provide the following models trained using the `CoNIC dataset <https://conic-challenge.grand-challenge.org/>`_, which uses the following
+input output configuration:
+
+.. collapse:: Input Output Configuration Details
+
+   .. code-block:: python
+
+        from tiatoolbox.models import IOPatchPredictorConfig
+        ioconfig = IOPatchPredictorConfig(
+            patch_input_shape=(31, 31),
+            stride_shape=(8, 8),
+            input_resolutions=[{"resolution": 0.25, "units": "mpp"}]
+        )
+
+
+.. collapse:: Model names
+
+    - sccnn-conic
+
+.. collapse:: Input Output Configuration Details
+
+   .. code-block:: python
+
+        from tiatoolbox.models import IOPatchPredictorConfig
+        ioconfig = IOPatchPredictorConfig(
+            patch_input_shape=(252, 252),
+            stride_shape=(150, 150),
+            input_resolutions=[{"resolution": 0.25, "units": "mpp"}]
+        )
+
+
+.. collapse:: Model names
+
+    - mapde-conic
+
+
 Multi-Task Segmentation
 ^^^^^^^^^^^^^^^^^^^^^^^
 
diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
new file mode 100644
index 000000000..bcb94bdb5
--- /dev/null
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -0,0 +1,8 @@
+"""Tests for NucleusDetector."""
+
+from tiatoolbox.models.engine.nucleus_detection import NucleusDetector
+
+
+def test_nucleus_detector_engine(sample_svs):
+    """Test for nucleus detection engine."""
+    _ = NucleusDetector(pretrained_model="mapde-conic")
diff --git a/tests/models/test_patch_predictor.py b/tests/models/test_patch_predictor.py
index 03d135cd8..465dd1f27 100644
--- a/tests/models/test_patch_predictor.py
+++ b/tests/models/test_patch_predictor.py
@@ -546,7 +546,7 @@ def test_io_config_delegation(remote_sample, tmp_path):
     predictor = PatchPredictor(pretrained_model="resnet18-kather100k", batch_size=1)
     predictor.predict(
         [mini_wsi_svs],
-        patch_input_shape=[300, 300],
+        patch_input_shape=(300, 300),
         mode="wsi",
         on_gpu=ON_GPU,
         save_dir=f"{tmp_path}/dump",
@@ -556,7 +556,7 @@ def test_io_config_delegation(remote_sample, tmp_path):
 
     predictor.predict(
         [mini_wsi_svs],
-        stride_shape=[300, 300],
+        stride_shape=(300, 300),
         mode="wsi",
         on_gpu=ON_GPU,
         save_dir=f"{tmp_path}/dump",
diff --git a/tiatoolbox/data/pretrained_model.yaml b/tiatoolbox/data/pretrained_model.yaml
index f9a480de5..2243cebb0 100644
--- a/tiatoolbox/data/pretrained_model.yaml
+++ b/tiatoolbox/data/pretrained_model.yaml
@@ -777,44 +777,66 @@ mapde-crchisto:
     architecture:
         class: mapde.MapDe
         kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.25 }
             num_input_channels: 3
             min_distance: 4
             threshold_abs: 250
             num_classes: 1
+            tile_shape: [ 2048, 2048 ]
+            patch_input_shape: [ 252, 252 ]
+            stride_shape: [ 150, 150 ]
 
 mapde-conic:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/mapde-conic.pth
     architecture:
         class: mapde.MapDe
         kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.25 }
             num_input_channels: 3
             min_distance: 3
             threshold_abs: 205
             num_classes: 1
+            tile_shape: [ 2048, 2048 ]
+            patch_input_shape: [ 252, 252 ]
+            stride_shape: [ 150, 150 ]
 
 sccnn-crchisto:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-crchisto.pth
     architecture:
         class: sccnn.SCCNN
         kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.25 }
             num_input_channels: 3
             out_height: 13
             out_width: 13
             radius: 12
             min_distance: 6
             threshold_abs: 0.20
+            tile_shape: [ 2048, 2048 ]
+            patch_input_shape: [ 31, 31 ]
+            patch_output_shape: [ 13, 13 ]
+            stride_shape: [ 8, 8 ]
 
 sccnn-conic:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-conic.pth
     architecture:
         class: sccnn.SCCNN
         kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.25 }
             num_input_channels: 3
             out_height: 13
             out_width: 13
             radius: 12
             min_distance: 5
             threshold_abs: 0.05
+            tile_shape: [ 2048, 2048 ]
+            patch_input_shape: [ 31, 31 ]
+            patch_output_shape: [ 13, 13 ]
+            stride_shape: [ 8, 8 ]
 
 nuclick_original-pannuke:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/seg/nuclick_original-pannuke.pth

From 46ffdae3ad2f980be3d189dd403655b3ad40a4a6 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Mon, 20 Feb 2023 17:33:47 +0000
Subject: [PATCH 04/14] :boom: DEV: Fix mpp for cell detection - Fix mpp for
 cell detection

---
 tiatoolbox/data/pretrained_model.yaml | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tiatoolbox/data/pretrained_model.yaml b/tiatoolbox/data/pretrained_model.yaml
index 2243cebb0..0f67004d8 100644
--- a/tiatoolbox/data/pretrained_model.yaml
+++ b/tiatoolbox/data/pretrained_model.yaml
@@ -778,7 +778,7 @@ mapde-crchisto:
         class: mapde.MapDe
         kwargs:
             input_resolutions:
-                - { "units": "mpp", "resolution": 0.25 }
+                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             min_distance: 4
             threshold_abs: 250
@@ -793,7 +793,7 @@ mapde-conic:
         class: mapde.MapDe
         kwargs:
             input_resolutions:
-                - { "units": "mpp", "resolution": 0.25 }
+                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             min_distance: 3
             threshold_abs: 205
@@ -808,7 +808,7 @@ sccnn-crchisto:
         class: sccnn.SCCNN
         kwargs:
             input_resolutions:
-                - { "units": "mpp", "resolution": 0.25 }
+                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             out_height: 13
             out_width: 13
@@ -826,7 +826,7 @@ sccnn-conic:
         class: sccnn.SCCNN
         kwargs:
             input_resolutions:
-                - { "units": "mpp", "resolution": 0.25 }
+                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             out_height: 13
             out_width: 13

From 4e2ca8bdd2622c65541f1660efa71508a69032d0 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 21 Feb 2023 14:18:49 +0000
Subject: [PATCH 05/14] :bug: BUG: Fix bug with loading NucleusDetector - Fix
 bug with loading NucleusDetector

---
 tests/models/test_nucleus_detection_engine.py |   2 +-
 tiatoolbox/data/pretrained_model.yaml         |  55 ++++++---
 tiatoolbox/models/engine/nucleus_detection.py | 109 ++++++++++++++----
 3 files changed, 127 insertions(+), 39 deletions(-)

diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
index bcb94bdb5..615873593 100644
--- a/tests/models/test_nucleus_detection_engine.py
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -5,4 +5,4 @@
 
 def test_nucleus_detector_engine(sample_svs):
     """Test for nucleus detection engine."""
-    _ = NucleusDetector(pretrained_model="mapde-conic")
+    _ = NucleusDetector(pretrained_model="mapde-conic", batch_size=1)
diff --git a/tiatoolbox/data/pretrained_model.yaml b/tiatoolbox/data/pretrained_model.yaml
index 0f67004d8..042f3aa5c 100644
--- a/tiatoolbox/data/pretrained_model.yaml
+++ b/tiatoolbox/data/pretrained_model.yaml
@@ -765,7 +765,6 @@ micronet-consep:
                 - {"units": "mpp", "resolution": 0.25}
             output_resolutions:
                 - {"units": "mpp", "resolution": 0.25}
-            margin: 128
             tile_shape: [2048, 2048]
             patch_input_shape: [252, 252]
             patch_output_shape: [252, 252]
@@ -777,30 +776,44 @@ mapde-crchisto:
     architecture:
         class: mapde.MapDe
         kwargs:
-            input_resolutions:
-                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             min_distance: 4
             threshold_abs: 250
             num_classes: 1
+    ioconfig:
+        class: semantic_segmentor.IOSegmentorConfig
+        kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
+            output_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
             patch_input_shape: [ 252, 252 ]
+            patch_output_shape: [ 252, 252 ]
             stride_shape: [ 150, 150 ]
+            save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 mapde-conic:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/mapde-conic.pth
     architecture:
         class: mapde.MapDe
         kwargs:
-            input_resolutions:
-                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
             min_distance: 3
             threshold_abs: 205
             num_classes: 1
+    ioconfig:
+        class: semantic_segmentor.IOSegmentorConfig
+        kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
+            output_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
             patch_input_shape: [ 252, 252 ]
+            patch_output_shape: [ 252, 252 ]
             stride_shape: [ 150, 150 ]
+            save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 sccnn-crchisto:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-crchisto.pth
@@ -814,11 +827,19 @@ sccnn-crchisto:
             out_width: 13
             radius: 12
             min_distance: 6
-            threshold_abs: 0.20
+            threshold_abs: 0.201
+    ioconfig:
+        class: semantic_segmentor.IOSegmentorConfig
+        kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
+            output_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
-            patch_input_shape: [ 31, 31 ]
-            patch_output_shape: [ 13, 13 ]
-            stride_shape: [ 8, 8 ]
+            patch_input_shape: [ 252, 252 ]
+            patch_output_shape: [ 252, 252 ]
+            stride_shape: [ 150, 150 ]
+            save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 sccnn-conic:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-conic.pth
@@ -832,11 +853,19 @@ sccnn-conic:
             out_width: 13
             radius: 12
             min_distance: 5
-            threshold_abs: 0.05
+            threshold_abs: 0.051
+    ioconfig:
+        class: semantic_segmentor.IOSegmentorConfig
+        kwargs:
+            input_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
+            output_resolutions:
+                - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
-            patch_input_shape: [ 31, 31 ]
-            patch_output_shape: [ 13, 13 ]
-            stride_shape: [ 8, 8 ]
+            patch_input_shape: [ 252, 252 ]
+            patch_output_shape: [ 252, 252 ]
+            stride_shape: [ 150, 150 ]
+            save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 nuclick_original-pannuke:
     url: https://tiatoolbox.dcs.warwick.ac.uk/models/seg/nuclick_original-pannuke.pth
diff --git a/tiatoolbox/models/engine/nucleus_detection.py b/tiatoolbox/models/engine/nucleus_detection.py
index b31e11d7d..5c35415ae 100644
--- a/tiatoolbox/models/engine/nucleus_detection.py
+++ b/tiatoolbox/models/engine/nucleus_detection.py
@@ -1,10 +1,66 @@
 """This module implements nucleus detection engine."""
 
 
-from tiatoolbox.models.engine.patch_predictor import PatchPredictor
+from typing import List, Union
 
+import numpy as np
 
-class NucleusDetector(PatchPredictor):
+from tiatoolbox.models.engine.semantic_segmentor import (
+    IOSegmentorConfig,
+    SemanticSegmentor,
+)
+
+
+class IONucleusDetectorConfig(IOSegmentorConfig):
+    """Contains NucleusDetector input and output information.
+
+    Args:
+        input_resolutions (list):
+            Resolution of each input head of model inference, must be in
+            the same order as `target model.forward()`.
+        output_resolutions (list):
+            Resolution of each output head from model inference, must be
+            in the same order as target model.infer_batch().
+        patch_input_shape (:class:`numpy.ndarray`, list(int)):
+            Shape of the largest input in (height, width).
+        patch_output_shape (:class:`numpy.ndarray`, list(int)):
+            Shape of the largest output in (height, width).
+        save_resolution (dict):
+            Resolution to save all output.
+
+    Examples:
+        >>> # Defining io for a network having 1 input and 1 output at the
+        >>> # same resolution
+        >>> ioconfig = IONucleusDetectorConfig(
+        ...     input_resolutions=[{"units": "baseline", "resolution": 1.0}],
+        ...     output_resolutions=[{"units": "baseline", "resolution": 1.0}],
+        ...     patch_input_shape=[2048, 2048],
+        ...     patch_output_shape=[1024, 1024],
+        ...     stride_shape=[512, 512],
+        ... )
+
+    """
+
+    def __init__(
+        self,
+        input_resolutions: List[dict],
+        output_resolutions: List[dict],
+        patch_input_shape: Union[List[int], np.ndarray],
+        patch_output_shape: Union[List[int], np.ndarray],
+        save_resolution: dict = None,
+        **kwargs,
+    ):
+        super().__init__(
+            input_resolutions=input_resolutions,
+            output_resolutions=output_resolutions,
+            patch_input_shape=patch_input_shape,
+            patch_output_shape=patch_output_shape,
+            save_resolution=save_resolution,
+            **kwargs,
+        )
+
+
+class NucleusDetector(SemanticSegmentor):
     r"""Nucleus detection engine.
 
     The models provided by tiatoolbox should give the following results:
@@ -31,8 +87,8 @@ class NucleusDetector(PatchPredictor):
             Path to the weight of the corresponding `pretrained_model`.
 
           >>> predictor = NucleusDetector(
-          ...    pretrained_model="resnet18-kather100k",
-          ...    pretrained_weights="resnet18_local_weight")
+          ...    pretrained_model="mapde-conic",
+          ...    pretrained_weights="mapde_local_weight")
 
         batch_size (int):
             Number of images fed into the model each time.
@@ -40,14 +96,14 @@ class NucleusDetector(PatchPredictor):
             Number of workers to load the data. Take note that they will
             also perform preprocessing.
         verbose (bool):
-            Whether to output logging information.
+            Whether to output logging information. default=False.
+        auto_generate_mask (bool):
+            To automatically generate tile/WSI tissue mask if is not
+            provided. default=False.
 
     Attributes:
         imgs (:obj:`str` or :obj:`pathlib.Path` or :obj:`numpy.ndarray`):
             A HWC image or a path to WSI.
-        mode (str):
-            Type of input to process. Choose from either `patch`, `tile`
-            or `wsi`.
         model (nn.Module):
             Defined PyTorch model.
         pretrained_model (str):
@@ -69,37 +125,38 @@ class NucleusDetector(PatchPredictor):
     Examples:
         >>> # list of 2 image patches as input
         >>> data = [img1, img2]
-        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
-        >>> output = predictor.predict(data, mode='patch')
+        >>> nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
+        >>> output = nucleus_detector.predict(data, mode='patch')
 
         >>> # array of list of 2 image patches as input
         >>> data = np.array([img1, img2])
-        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
-        >>> output = predictor.predict(data, mode='patch')
+        >>> nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
+        >>> output = nucleus_detector.predict(data, mode='patch')
 
         >>> # list of 2 image patch files as input
         >>> data = ['path/img.png', 'path/img.png']
-        >>> predictor = PatchPredictor(pretrained_model="resnet18-kather100k")
-        >>> output = predictor.predict(data, mode='patch')
+        >>> nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
+        >>> output = nucleus_detector.predict(data, mode='patch')
 
         >>> # list of 2 image tile files as input
         >>> tile_file = ['path/tile1.png', 'path/tile2.png']
-        >>> predictor = PatchPredictor(pretraind_model="resnet18-kather100k")
-        >>> output = predictor.predict(tile_file, mode='tile')
+        >>> nucleus_detector = NucleusDetector(pretraind_model="mapde-conic")
+        >>> output = nucleus_detector.predict(tile_file, mode='tile')
 
         >>> # list of 2 wsi files as input
         >>> wsi_file = ['path/wsi1.svs', 'path/wsi2.svs']
-        >>> predictor = PatchPredictor(pretraind_model="resnet18-kather100k")
-        >>> output = predictor.predict(wsi_file, mode='wsi')
+        >>> nucleus_detector = NucleusDetector(pretraind_model="mapde-conic")
+        >>> output = nucleus_detector.predict(wsi_file, mode='wsi')
 
     References:
-        [1] Kather, Jakob Nikolas, et al. "Predicting survival from colorectal cancer
-        histology slides using deep learning: A retrospective multicenter study."
-        PLoS medicine 16.1 (2019): e1002730.
+        [1] Raza, Shan E. Ahmed, et al. "Deconvolving convolutional neural network
+        for cell detection." 2019 IEEE 16th International Symposium on Biomedical
+        Imaging (ISBI 2019). IEEE, 2019.
 
-        [2] Veeling, Bastiaan S., et al. "Rotation equivariant CNNs for digital
-        pathology." International Conference on Medical image computing and
-        computer-assisted intervention. Springer, Cham, 2018.
+        [2] Sirinukunwattana, Korsuk, et al.
+        "Locality sensitive deep learning for detection and classification
+        of nuclei in routine colon cancer histology images."
+        IEEE transactions on medical imaging 35.5 (2016): 1196-1206.
 
     """  # noqa: W605
 
@@ -110,7 +167,8 @@ def __init__(
         model=None,
         pretrained_model=None,
         pretrained_weights=None,
-        verbose=True,
+        verbose: bool = False,
+        auto_generate_mask: bool = False,
     ):
         super().__init__(
             batch_size=batch_size,
@@ -119,4 +177,5 @@ def __init__(
             pretrained_model=pretrained_model,
             pretrained_weights=pretrained_weights,
             verbose=verbose,
+            auto_generate_mask=auto_generate_mask,
         )

From 1af81403acf9a62ea2bc72c3fd8ccb1ce698c38e Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 21 Feb 2023 14:33:38 +0000
Subject: [PATCH 06/14] :bug: BUG: Fix bug with loading sccnn - Fix bug with
 loading sccnn

---
 tests/models/test_nucleus_detection_engine.py |  2 +-
 tiatoolbox/data/pretrained_model.yaml         | 26 +++++++------------
 2 files changed, 11 insertions(+), 17 deletions(-)

diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
index 615873593..0e7346e72 100644
--- a/tests/models/test_nucleus_detection_engine.py
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -5,4 +5,4 @@
 
 def test_nucleus_detector_engine(sample_svs):
     """Test for nucleus detection engine."""
-    _ = NucleusDetector(pretrained_model="mapde-conic", batch_size=1)
+    _ = NucleusDetector(pretrained_model="sccnn-crchisto")
diff --git a/tiatoolbox/data/pretrained_model.yaml b/tiatoolbox/data/pretrained_model.yaml
index 042f3aa5c..9eb539efc 100644
--- a/tiatoolbox/data/pretrained_model.yaml
+++ b/tiatoolbox/data/pretrained_model.yaml
@@ -820,14 +820,11 @@ sccnn-crchisto:
     architecture:
         class: sccnn.SCCNN
         kwargs:
-            input_resolutions:
-                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
-            out_height: 13
-            out_width: 13
             radius: 12
             min_distance: 6
-            threshold_abs: 0.201
+            threshold_abs: 0.20
+            patch_output_shape: [ 13, 13 ]
     ioconfig:
         class: semantic_segmentor.IOSegmentorConfig
         kwargs:
@@ -836,9 +833,9 @@ sccnn-crchisto:
             output_resolutions:
                 - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
-            patch_input_shape: [ 252, 252 ]
-            patch_output_shape: [ 252, 252 ]
-            stride_shape: [ 150, 150 ]
+            patch_input_shape: [ 31, 31 ]
+            patch_output_shape: [ 13, 13 ]
+            stride_shape: [ 8, 8 ]
             save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 sccnn-conic:
@@ -846,14 +843,11 @@ sccnn-conic:
     architecture:
         class: sccnn.SCCNN
         kwargs:
-            input_resolutions:
-                - { "units": "mpp", "resolution": 0.5 }
             num_input_channels: 3
-            out_height: 13
-            out_width: 13
             radius: 12
             min_distance: 5
-            threshold_abs: 0.051
+            threshold_abs: 0.05
+            patch_output_shape: [ 13, 13 ]
     ioconfig:
         class: semantic_segmentor.IOSegmentorConfig
         kwargs:
@@ -862,9 +856,9 @@ sccnn-conic:
             output_resolutions:
                 - { "units": "mpp", "resolution": 0.5 }
             tile_shape: [ 2048, 2048 ]
-            patch_input_shape: [ 252, 252 ]
-            patch_output_shape: [ 252, 252 ]
-            stride_shape: [ 150, 150 ]
+            patch_input_shape: [ 31, 31 ]
+            patch_output_shape: [ 13, 13 ]
+            stride_shape: [ 8, 8 ]
             save_resolution: { 'units': 'mpp', 'resolution': 0.5 }
 
 nuclick_original-pannuke:

From 527444a7da96a9ff263addee043c87d9ded4de89 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 21 Feb 2023 15:05:25 +0000
Subject: [PATCH 07/14] :bug: BUG: Fix test bug in test_patch_predictor.py -
 Fix test bug in test_patch_predictor.py

---
 tests/models/test_patch_predictor.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tests/models/test_patch_predictor.py b/tests/models/test_patch_predictor.py
index 465dd1f27..09ab9a9ab 100644
--- a/tests/models/test_patch_predictor.py
+++ b/tests/models/test_patch_predictor.py
@@ -551,7 +551,7 @@ def test_io_config_delegation(remote_sample, tmp_path):
         on_gpu=ON_GPU,
         save_dir=f"{tmp_path}/dump",
     )
-    assert predictor._ioconfig.patch_input_shape == [300, 300]
+    assert predictor._ioconfig.patch_input_shape == (300, 300)
     _rm_dir(f"{tmp_path}/dump")
 
     predictor.predict(
@@ -561,7 +561,7 @@ def test_io_config_delegation(remote_sample, tmp_path):
         on_gpu=ON_GPU,
         save_dir=f"{tmp_path}/dump",
     )
-    assert predictor._ioconfig.stride_shape == [300, 300]
+    assert predictor._ioconfig.stride_shape == (300, 300)
     _rm_dir(f"{tmp_path}/dump")
 
     predictor.predict(

From 37b82f9153d17cc8e97d6d1c62b703a7c7b23b4c Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 21 Feb 2023 17:41:19 +0000
Subject: [PATCH 08/14] :white_check_mark: TST: Add test for NucleusDetector -
 Add test for NucleusDetector

---
 tests/models/test_nucleus_detection_engine.py | 22 +++++-
 tiatoolbox/models/engine/__init__.py          |  2 +-
 ...cleus_detection.py => nucleus_detector.py} | 72 +++++++++++++++++++
 3 files changed, 92 insertions(+), 4 deletions(-)
 rename tiatoolbox/models/engine/{nucleus_detection.py => nucleus_detector.py} (68%)

diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
index 0e7346e72..762e4ed41 100644
--- a/tests/models/test_nucleus_detection_engine.py
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -1,8 +1,24 @@
 """Tests for NucleusDetector."""
 
-from tiatoolbox.models.engine.nucleus_detection import NucleusDetector
+import pathlib
 
+import pandas as pd
+import pytest
 
-def test_nucleus_detector_engine(sample_svs):
+from tiatoolbox.models.engine.nucleus_detector import NucleusDetector
+
+
+def test_nucleus_detector_engine(remote_sample, tmp_path):
     """Test for nucleus detection engine."""
-    _ = NucleusDetector(pretrained_model="sccnn-crchisto")
+    mini_wsi_svs = pathlib.Path(remote_sample("wsi4_512_512_svs"))
+
+    nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
+    _ = nucleus_detector.predict(
+        [mini_wsi_svs], mode="wsi", save_dir=tmp_path / "output"
+    )
+
+    coordinates = pd.read_csv(tmp_path / "output" / "0.locations.0.csv")
+    assert coordinates.x[0] == pytest.approx(53, abs=2)
+    assert coordinates.x[1] == pytest.approx(55, abs=2)
+    assert coordinates.y[0] == pytest.approx(107, abs=2)
+    assert coordinates.y[1] == pytest.approx(127, abs=2)
diff --git a/tiatoolbox/models/engine/__init__.py b/tiatoolbox/models/engine/__init__.py
index 7574d6e7e..e7936a1c1 100644
--- a/tiatoolbox/models/engine/__init__.py
+++ b/tiatoolbox/models/engine/__init__.py
@@ -1,6 +1,6 @@
 """Engines to run models implemented in tiatoolbox."""
 from tiatoolbox.models.engine import (
-    nucleus_detection,
+    nucleus_detector,
     nucleus_instance_segmentor,
     patch_predictor,
     semantic_segmentor,
diff --git a/tiatoolbox/models/engine/nucleus_detection.py b/tiatoolbox/models/engine/nucleus_detector.py
similarity index 68%
rename from tiatoolbox/models/engine/nucleus_detection.py
rename to tiatoolbox/models/engine/nucleus_detector.py
index 5c35415ae..9eed8e324 100644
--- a/tiatoolbox/models/engine/nucleus_detection.py
+++ b/tiatoolbox/models/engine/nucleus_detector.py
@@ -4,6 +4,7 @@
 from typing import List, Union
 
 import numpy as np
+import pandas as pd
 
 from tiatoolbox.models.engine.semantic_segmentor import (
     IOSegmentorConfig,
@@ -160,6 +161,8 @@ class NucleusDetector(SemanticSegmentor):
 
     """  # noqa: W605
 
+    from tiatoolbox.wsicore.wsireader import WSIReader
+
     def __init__(
         self,
         batch_size=8,
@@ -179,3 +182,72 @@ def __init__(
             verbose=verbose,
             auto_generate_mask=auto_generate_mask,
         )
+
+    def _process_predictions(
+        self,
+        cum_batch_predictions: List,
+        wsi_reader: WSIReader,
+        ioconfig: IOSegmentorConfig,
+        save_path: str,
+        cache_dir: str,
+    ):
+        """Define how the aggregated predictions are processed.
+
+        This includes merging the prediction if necessary and also saving the
+        locations afterwards. Note that items within `cum_batch_predictions` will
+        be consumed during the operation.
+
+        Args:
+            cum_batch_predictions (list):
+                List of batch predictions. Each item within the list
+                should be of (location, patch_predictions).
+            wsi_reader (:class:`WSIReader`):
+                A reader for the image where the predictions come from.
+            ioconfig (:class:`IOSegmentorConfig`):
+                A configuration object contains input and output
+                information.
+            save_path (str):
+                Root path to save current WSI predictions.
+            cache_dir (str):
+                Root path to cache current WSI data.
+
+        """
+        if len(cum_batch_predictions) == 0:
+            return
+
+        # assume predictions is N, each item has L output element
+        locations, predictions = list(zip(*cum_batch_predictions))
+        # Nx4 (N x [tl_x, tl_y, br_x, br_y), denotes the location of
+        # output patch this can exceed the image bound at the requested
+        # resolution remove singleton due to split.
+        locations = np.array([v[0] for v in locations])
+        for index, output_resolution in enumerate(ioconfig.output_resolutions):
+            # assume resolution index to be in the same order as L
+            merged_resolution = ioconfig.highest_input_resolution
+            merged_locations = locations
+            # ! location is w.r.t the highest resolution, hence still need conversion
+            if ioconfig.save_resolution is not None:
+                merged_resolution = ioconfig.save_resolution
+                output_shape = wsi_reader.slide_dimensions(**output_resolution)
+                merged_shape = wsi_reader.slide_dimensions(**merged_resolution)
+                fx = merged_shape[0] / output_shape[0]
+                merged_locations = np.ceil(locations * fx).astype(np.int64)
+            merged_shape = wsi_reader.slide_dimensions(**merged_resolution)
+            # 0 idx is to remove singleton without removing other axes singleton
+            to_merge_predictions = [v[index][0] for v in predictions]
+            sub_save_path = f"{save_path}.raw.{index}.npy"
+            sub_count_path = f"{cache_dir}/count.{index}.npy"
+            cum_canvas = self.merge_prediction(
+                merged_shape[::-1],  # XY to YX
+                to_merge_predictions,
+                merged_locations,
+                save_path=sub_save_path,
+                cache_count_path=sub_count_path,
+            )
+
+            # Coordinates in output resolution for the current canvas.
+            cum_canvas = np.expand_dims(cum_canvas, axis=0)
+            coordinates_canvas = pd.DataFrame(
+                self.model.postproc_func(cum_canvas), columns=["x", "y"]
+            )
+            coordinates_canvas.to_csv(f"{save_path}.locations.{index}.csv", index=False)

From 1405fa0a93aec61743b53628810d58621967641e Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 21 Feb 2023 18:26:51 +0000
Subject: [PATCH 09/14] :bug: TST: Set GPU - Set GPU

Signed-off-by: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
---
 tests/models/test_nucleus_detection_engine.py  | 8 +++++++-
 tiatoolbox/models/engine/semantic_segmentor.py | 2 +-
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
index 762e4ed41..a6906720e 100644
--- a/tests/models/test_nucleus_detection_engine.py
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -6,6 +6,9 @@
 import pytest
 
 from tiatoolbox.models.engine.nucleus_detector import NucleusDetector
+from tiatoolbox.utils import env_detection as toolbox_env
+
+ON_GPU = not toolbox_env.running_on_ci() and toolbox_env.has_gpu()
 
 
 def test_nucleus_detector_engine(remote_sample, tmp_path):
@@ -14,7 +17,10 @@ def test_nucleus_detector_engine(remote_sample, tmp_path):
 
     nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
     _ = nucleus_detector.predict(
-        [mini_wsi_svs], mode="wsi", save_dir=tmp_path / "output"
+        [mini_wsi_svs],
+        mode="wsi",
+        save_dir=tmp_path / "output",
+        on_gpu=ON_GPU,
     )
 
     coordinates = pd.read_csv(tmp_path / "output" / "0.locations.0.csv")
diff --git a/tiatoolbox/models/engine/semantic_segmentor.py b/tiatoolbox/models/engine/semantic_segmentor.py
index 7fe749d49..1edf2b92d 100644
--- a/tiatoolbox/models/engine/semantic_segmentor.py
+++ b/tiatoolbox/models/engine/semantic_segmentor.py
@@ -1531,7 +1531,7 @@ def predict(
             units (str):
                 Units of resolution used for reading the image. Choose
                 from either `"level"`, `"power"` or `"mpp"`.
-            save_dir (str):
+            save_dir (str or pathlib.Path):
                 Output directory when processing multiple tiles and
                 whole-slide images. By default, it is folder `output`
                 where the running script is invoked.

From 71cb5c6fd05e8f58aac5e8395290c9d8836d79de Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Wed, 22 Feb 2023 09:50:10 +0000
Subject: [PATCH 10/14] :white_check_mark: TST: Add test to improve coverage -
 Add test to improve coverage

---
 tests/models/test_nucleus_detection_engine.py | 49 ++++++++++++++++---
 1 file changed, 43 insertions(+), 6 deletions(-)

diff --git a/tests/models/test_nucleus_detection_engine.py b/tests/models/test_nucleus_detection_engine.py
index a6906720e..e9f4e1aec 100644
--- a/tests/models/test_nucleus_detection_engine.py
+++ b/tests/models/test_nucleus_detection_engine.py
@@ -1,16 +1,35 @@
 """Tests for NucleusDetector."""
 
 import pathlib
+import shutil
 
 import pandas as pd
 import pytest
 
-from tiatoolbox.models.engine.nucleus_detector import NucleusDetector
+from tiatoolbox.models.engine.nucleus_detector import (
+    IONucleusDetectorConfig,
+    NucleusDetector,
+)
 from tiatoolbox.utils import env_detection as toolbox_env
 
 ON_GPU = not toolbox_env.running_on_ci() and toolbox_env.has_gpu()
 
 
+def _rm_dir(path):
+    """Helper func to remove directory."""
+    if pathlib.Path(path).exists():
+        shutil.rmtree(path, ignore_errors=True)
+
+
+def check_output(path):
+    """Check NucleusDetector output."""
+    coordinates = pd.read_csv(path)
+    assert coordinates.x[0] == pytest.approx(53, abs=2)
+    assert coordinates.x[1] == pytest.approx(55, abs=2)
+    assert coordinates.y[0] == pytest.approx(107, abs=2)
+    assert coordinates.y[1] == pytest.approx(127, abs=2)
+
+
 def test_nucleus_detector_engine(remote_sample, tmp_path):
     """Test for nucleus detection engine."""
     mini_wsi_svs = pathlib.Path(remote_sample("wsi4_512_512_svs"))
@@ -23,8 +42,26 @@ def test_nucleus_detector_engine(remote_sample, tmp_path):
         on_gpu=ON_GPU,
     )
 
-    coordinates = pd.read_csv(tmp_path / "output" / "0.locations.0.csv")
-    assert coordinates.x[0] == pytest.approx(53, abs=2)
-    assert coordinates.x[1] == pytest.approx(55, abs=2)
-    assert coordinates.y[0] == pytest.approx(107, abs=2)
-    assert coordinates.y[1] == pytest.approx(127, abs=2)
+    check_output(tmp_path / "output" / "0.locations.0.csv")
+
+    _rm_dir(tmp_path / "output")
+
+    ioconfig = IONucleusDetectorConfig(
+        input_resolutions=[{"units": "mpp", "resolution": 0.5}],
+        output_resolutions=[{"units": "mpp", "resolution": 0.5}],
+        save_resolution=None,
+        patch_input_shape=[252, 252],
+        patch_output_shape=[252, 252],
+        stride_shape=[150, 150],
+    )
+
+    nucleus_detector = NucleusDetector(pretrained_model="mapde-conic")
+    _ = nucleus_detector.predict(
+        [mini_wsi_svs],
+        mode="wsi",
+        save_dir=tmp_path / "output",
+        on_gpu=ON_GPU,
+        ioconfig=ioconfig,
+    )
+
+    check_output(tmp_path / "output" / "0.locations.0.csv")

From b11e6f0eeaf2aa85b26a4c802da3cb707e1c2398 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Tue, 2 May 2023 16:51:02 +0100
Subject: [PATCH 11/14] :memo: Add example models - Add example models

---
 tiatoolbox/models/engine/nucleus_detector.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tiatoolbox/models/engine/nucleus_detector.py b/tiatoolbox/models/engine/nucleus_detector.py
index 9eed8e324..6e8e8615b 100644
--- a/tiatoolbox/models/engine/nucleus_detector.py
+++ b/tiatoolbox/models/engine/nucleus_detector.py
@@ -109,8 +109,8 @@ class NucleusDetector(SemanticSegmentor):
             Defined PyTorch model.
         pretrained_model (str):
             Name of the existing models support by tiatoolbox for
-            processing the data. For a full list of pretrained models,
-            refer to the `docs
+            processing the data e.g., mapde-conic, sccnn-conic.
+            For a full list of pretrained models, please refer to the `docs
             <https://tia-toolbox.readthedocs.io/en/latest/pretrained.html>`_
             By default, the corresponding pretrained weights will also
             be downloaded. However, you can override with your own set

From 2c3884a826a9d21eafd91a40a902723b75648cdf Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Fri, 2 Jun 2023 10:59:10 +0100
Subject: [PATCH 12/14] :pushpin: Pin `torch` version - Pin `torch` version

---
 requirements/requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements/requirements.txt b/requirements/requirements.txt
index 8195c2f32..532b4c182 100644
--- a/requirements/requirements.txt
+++ b/requirements/requirements.txt
@@ -25,7 +25,7 @@ shapely>=2.0.0
 SimpleITK>=2.2.1
 sphinx>=5.3.0
 tifffile>=2022.10.10
-torch>=2.0.0
+torch>=1.13.0, <2.0.0
 torchvision>=0.14.1
 tqdm>=4.64.1
 umap-learn>=0.5.3

From e670455da612fa82b3929655143266aad1a74052 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Fri, 2 Jun 2023 11:41:53 +0100
Subject: [PATCH 13/14] :pushpin: Pin `torch` version - Pin `torch` version

---
 requirements/requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements/requirements.txt b/requirements/requirements.txt
index 532b4c182..22242dc86 100644
--- a/requirements/requirements.txt
+++ b/requirements/requirements.txt
@@ -25,7 +25,7 @@ shapely>=2.0.0
 SimpleITK>=2.2.1
 sphinx>=5.3.0
 tifffile>=2022.10.10
-torch>=1.13.0, <2.0.0
+torch>=1.13.0
 torchvision>=0.14.1
 tqdm>=4.64.1
 umap-learn>=0.5.3

From 7989aa7b06cfdfc186c40a35421155e0dc49be64 Mon Sep 17 00:00:00 2001
From: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
Date: Sat, 17 Jun 2023 14:57:57 +0100
Subject: [PATCH 14/14] :hammer: Try cuda 11.8 - Try cuda 11.8 to run the tests
 as it passes locally.

Signed-off-by: Shan E Ahmed Raza <13048456+shaneahmed@users.noreply.github.com>
---
 requirements/requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements/requirements.txt b/requirements/requirements.txt
index 22242dc86..719d9551a 100644
--- a/requirements/requirements.txt
+++ b/requirements/requirements.txt
@@ -1,5 +1,5 @@
 # torch installation
---extra-index-url https://download.pytorch.org/whl/cu117; sys_platform != "darwin"
+--extra-index-url https://download.pytorch.org/whl/cu118; sys_platform != "darwin"
 albumentations>=1.3.0
 Click>=8.1.3
 defusedxml>=0.7.1