NEW: Add SCCNN Cell Detection Architecture

tests/models/test_arch_micronet.py

@@ -1,5 +1,5 @@
 # skipcq: PTC-W6004
-"""Unit test package for HoVerNet."""
+"""Unit test package for MicroNet."""
 
 import pathlib
 

tests/models/test_arch_sccnn.py

@@ -0,0 +1,46 @@
+"""Unit test package for SCCNN."""
+import numpy as np
+import torch
+
+from tiatoolbox import utils
+from tiatoolbox.models.architecture import fetch_pretrained_weights
+from tiatoolbox.models.architecture.sccnn import SCCNN
+from tiatoolbox.wsicore.wsireader import WSIReader
+
+
+def _load_sccnn(tmp_path, name):
+    """Loads SCCNN model with specified weights."""
+    model = SCCNN()
+    fetch_pretrained_weights(name, f"{tmp_path}/weights.pth")
+    map_location = utils.misc.select_device(utils.env_detection.has_gpu())
+    pretrained = torch.load(f"{tmp_path}/weights.pth", map_location=map_location)
+    model.load_state_dict(pretrained)
+
+    return model
+
+
+def test_functionality(remote_sample, tmp_path):
+    """Functionality test for SCCNN.
+
+    Tests the functionality of SCCNN model for inference at the patch level.
+
+    """
+    tmp_path = str(tmp_path)
+    sample_wsi = str(remote_sample("wsi1_2k_2k_svs"))
+    reader = WSIReader.open(sample_wsi)
+
+    # * test fast mode (architecture used in PanNuke paper)
+    patch = reader.read_bounds(
+        (30, 30, 61, 61), resolution=0.25, units="mpp", coord_space="resolution"
+    )
+    patch = SCCNN.preproc(patch)
+    batch = torch.from_numpy(patch)[None]
+    model = _load_sccnn(tmp_path=tmp_path, name="sccnn-crchisto")
+    output = model.infer_batch(model, batch, on_gpu=False)
+    output = model.postproc(output[0])
+    assert np.all(output == [[8, 7]])
+
+    model = _load_sccnn(tmp_path=tmp_path, name="sccnn-conic")
+    output = model.infer_batch(model, batch, on_gpu=False)
+    output = model.postproc(output[0])
+    assert np.all(output == [[7, 8]])

tiatoolbox/data/pretrained_model.yaml

@@ -746,3 +746,27 @@ micronet-consep:
             patch_output_shape: [252, 252]
             stride_shape: [150, 150]
             save_resolution: {'units': 'mpp', 'resolution': 0.25}
+
+sccnn-crchisto:
+    url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-crchisto.pth
+    architecture:
+        class: sccnn.SCCNN
+        kwargs:
+            num_input_channels: 3
+            out_height: 13
+            out_width: 13
+            radius: 12
+            min_distance: 6
+            threshold_abs: 0.20
+
+sccnn-conic:
+    url: https://tiatoolbox.dcs.warwick.ac.uk/models/detection/sccnn-conic.pth
+    architecture:
+        class: sccnn.SCCNN
+        kwargs:
+            num_input_channels: 3
+            out_height: 13
+            out_width: 13
+            radius: 12
+            min_distance: 5
+            threshold_abs: 0.05

tiatoolbox/models/architecture/micronet.py

@@ -343,9 +343,10 @@ def out_arch_branch(in_ch, num_class=2):
 
 
 class MicroNet(ModelABC):
-    """Initialise MicroNet [1].
+    """Initialize MicroNet [1].
 
     The following models have been included in tiatoolbox:
+
     1. `micronet-consep`:
         This is trained on `CoNSeP dataset
         <https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/>`_ The