histolab · alessiamarcolini · Oct 13, 2021 · Oct 8, 2021 · Oct 8, 2021 · Oct 8, 2021
diff --git a/histolab/filters/image_filters_functional.py b/histolab/filters/image_filters_functional.py
@@ -476,13 +476,13 @@ def rgb_to_hed(img: PIL.Image.Image) -> PIL.Image.Image:
     if img.mode not in ["RGB", "RGBA"]:
         raise Exception("Input image must be RGB.")
     if img.mode == "RGBA":
-        img_arr = np.array(sk_color.rgba2rgb(img))
+        red, green, blue, _ = img.split()
+        img = PIL.Image.merge("RGB", (red, green, blue))
         warn(
             "Input image must be RGB. "
             "NOTE: the image will be converted to RGB before HED conversion."
         )
-    else:
-        img_arr = np.array(img)
+    img_arr = np.array(img)
     hed_arr = sk_color.rgb2hed(img_arr)
     hed = np_to_pil(hed_arr)
 

diff --git a/tests/expectations/pil-images-gs/diagnostic-slide-thumb-gs-kmeans-segmentation.png b/tests/expectations/pil-images-gs/diagnostic-slide-thumb-gs-kmeans-segmentation.png
diff --git a/tests/expectations/pil-images-gs/diagnostic-slide-thumb-gs-rag-threshold.png b/tests/expectations/pil-images-gs/diagnostic-slide-thumb-gs-rag-threshold.png
diff --git a/...s/expectations/pil-images-rgb/diagnostic-slide-thumb-gs-kmeans-segmentation.png b/...s/expectations/pil-images-rgb/diagnostic-slide-thumb-gs-kmeans-segmentation.png
diff --git a/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-gs-rag-threshold.png b/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-gs-rag-threshold.png
diff --git a/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-hsv-rgb-to-hed.png b/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-hsv-rgb-to-hed.png
diff --git a/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-rgb-to-hed.png b/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-rgb-to-hed.png
diff --git a/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-ycbcr-rgb-to-hed.png b/tests/expectations/pil-images-rgb/diagnostic-slide-thumb-ycbcr-rgb-to-hed.png
diff --git a/tests/expectations/pil-images-rgb/tcga-lung-rgb-dab-channel.png b/tests/expectations/pil-images-rgb/tcga-lung-rgb-dab-channel.png
diff --git a/tests/expectations/pil-images-rgb/tcga-lung-rgb-eosin-channel.png b/tests/expectations/pil-images-rgb/tcga-lung-rgb-eosin-channel.png
diff --git a/tests/expectations/pil-images-rgb/tcga-lung-rgb-hematoxylin-channel.png b/tests/expectations/pil-images-rgb/tcga-lung-rgb-hematoxylin-channel.png
diff --git a/tests/expectations/pil-images-rgba/diagnostic-slide-thumb-rgba-to-hed.png b/tests/expectations/pil-images-rgba/diagnostic-slide-thumb-rgba-to-hed.png
diff --git a/tests/expectations/pil-images-rgba/tcga-lung-dab-channel.png b/tests/expectations/pil-images-rgba/tcga-lung-dab-channel.png
diff --git a/tests/expectations/pil-images-rgba/tcga-lung-eosin-channel.png b/tests/expectations/pil-images-rgba/tcga-lung-eosin-channel.png
diff --git a/tests/expectations/pil-images-rgba/tcga-lung-hematoxylin-channel.png b/tests/expectations/pil-images-rgba/tcga-lung-hematoxylin-channel.png
diff --git a/...xpectations/tiles-location-images/cmu-1-small-region-tl-score-false-512x512.png b/...xpectations/tiles-location-images/cmu-1-small-region-tl-score-false-512x512.png
diff --git a/...expectations/tiles-location-images/cmu-1-small-region-tl-score-true-120x128.png b/...expectations/tiles-location-images/cmu-1-small-region-tl-score-true-120x128.png
diff --git a/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-f-512x512.png b/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-f-512x512.png
diff --git a/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-f-512x530.png b/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-f-512x530.png
diff --git a/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-t-128x128.png b/...pectations/tiles-location-images/tcga-cr-7395-01a-01-ts1-tl-score-t-128x128.png
diff --git a/tests/integration/test_image_filters.py b/tests/integration/test_image_filters.py
@@ -71,27 +71,12 @@ def test_invert_filter_with_gs_image():
     )
 
 
-@pytest.mark.parametrize(
-    "pil_image, expected_image",
-    (
-        (
-            RGB.DIAGNOSTIC_SLIDE_THUMB_RGB,
-            "pil-images-rgb/diagnostic-slide-thumb-rgb-to-hed",
-        ),
-        (
-            RGB.DIAGNOSTIC_SLIDE_THUMB_HSV,
-            "pil-images-rgb/diagnostic-slide-thumb-hsv-rgb-to-hed",
-        ),
-        (
-            RGB.DIAGNOSTIC_SLIDE_THUMB_YCBCR,
-            "pil-images-rgb/diagnostic-slide-thumb-ycbcr-rgb-to-hed",
-        ),
-    ),
-)
-def test_rgb_to_hed_filter_with_rgb_image(pil_image, expected_image):
-    expected_value = load_expectation(expected_image, type_="png")
+def test_rgb_to_hed_filter_with_rgb_image():
+    expected_value = load_expectation(
+        "pil-images-rgb/diagnostic-slide-thumb-rgb-to-hed", type_="png"
+    )
 
-    hed_img = imf.rgb_to_hed(pil_image)
+    hed_img = imf.rgb_to_hed(RGB.DIAGNOSTIC_SLIDE_THUMB_RGB)
 
     np.testing.assert_array_almost_equal(np.array(hed_img), np.array(expected_value))
     assert np.unique(np.array(ImageChops.difference(hed_img, expected_value)))[0] == 0
@@ -509,25 +494,6 @@ def test_local_equalization_raises_exception_on_rgb_images(pil_rgb_image):
     assert str(err.value) == "Input must be 2D."
 
 
-def test_kmeans_segmentation_filter_on_rgba_image():
-    rgba_img = RGBA.DIAGNOSTIC_SLIDE_THUMB
-    expected_value = load_expectation(
-        "pil-images-rgba/diagnostic-slide-thumb-kmeans-segmentation", type_="png"
-    )
-
-    kmeans_segmentation_img = imf.kmeans_segmentation(rgba_img, 800, 10)
-
-    np.testing.assert_array_almost_equal(
-        np.array(kmeans_segmentation_img), np.array(expected_value)
-    )
-    assert (
-        np.unique(
-            np.array(ImageChops.difference(kmeans_segmentation_img, expected_value))
-        )[0]
-        == 0
-    )
-
-
 def test_kmeans_segmentation_raises_value_error_on_rgba_images():
     rgba_img = RGBA.DIAGNOSTIC_SLIDE_THUMB
 
@@ -574,7 +540,7 @@ def test_kmeans_segmentation_filter_on_rgb_image(pil_image, expected_image):
 def test_kmeans_segmentation_filter_on_gs_image():
     gs_img = GS.DIAGNOSTIC_SLIDE_THUMB_GS
     expected_value = load_expectation(
-        "pil-images-gs/diagnostic-slide-thumb-gs-kmeans-segmentation", type_="png"
+        "pil-images-rgb/diagnostic-slide-thumb-gs-kmeans-segmentation", type_="png"
     )
 
     kmeans_segmentation_img = imf.kmeans_segmentation(gs_img, 800, 10)
@@ -624,7 +590,7 @@ def test_rag_threshold_filter_on_rgb_image(pil_image, expected_image):
 def test_rag_threshold_filter_on_gs_image():
     gs_img = GS.DIAGNOSTIC_SLIDE_THUMB_GS
     expected_value = load_expectation(
-        "pil-images-gs/diagnostic-slide-thumb-gs-rag-threshold", type_="png"
+        "pil-images-rgb/diagnostic-slide-thumb-gs-rag-threshold", type_="png"
     )
 
     rag_threshold_img = imf.rag_threshold(gs_img, 650, 20.6, 15)

diff --git a/tests/integration/test_scorer.py b/tests/integration/test_scorer.py
@@ -1,7 +1,7 @@
 import pytest
+from histolab.tile import Tile
 
 from histolab import scorer
-from histolab.tile import Tile
 
 from ..fixtures import TILES
 
@@ -11,38 +11,30 @@ class Describe_Scorers:
         "tile_img, expected_score",
         (
             # level 0
-            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, 4.95387907194613e-05),
-            (TILES.LOW_NUCLEI_SCORE_LEVEL0, 0.011112025501054716),
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, 0.018651677436394662),
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, 0.39901978131493154),
+            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, 0.0042),
+            (TILES.LOW_NUCLEI_SCORE_LEVEL0, 0.03001),
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, 0.266),
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, 0.008),
             # level 1
             (
                 TILES.VERY_LOW_NUCLEI_SCORE_RED_PEN_LEVEL1,
-                0.0017590279896743531,
-            ),  # breast - red pen
-            (TILES.LOW_NUCLEI_SCORE_LEVEL1, 0.019689596845556157),  # breast - green pen
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, 0.009512701556682022),  # aorta
-            (
-                TILES.MEDIUM_NUCLEI_SCORE_LEVEL1_2,
-                0.1519627864167197,
-            ),  # breast - green pen
-            (
-                TILES.MEDIUM_NUCLEI_SCORE_GREEN_PEN_LEVEL1,
-                0.35696740368342295,
-            ),  # breast - green pen
-            (
-                TILES.HIGH_NUCLEI_SCORE_RED_PEN_LEVEL1,
-                0.19867406253648537,
+                0.10605,
             ),  # breast - red pen
+            (TILES.LOW_NUCLEI_SCORE_LEVEL1, 0.00704),  # breast - green pen
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, 0.08904),  # aorta
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1_2, 0.09247),  # breast - green pen
+            (TILES.MEDIUM_NUCLEI_SCORE_GREEN_PEN_LEVEL1, 0.08088),
+            # breast - green pen
+            (TILES.HIGH_NUCLEI_SCORE_RED_PEN_LEVEL1, 0.10297),  # breast - red pen
             # level 2
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, 0.17959041877765877),  # prostate
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, 0.02165773137397028),  # prostate
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, 0.00337),  # prostate
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, 0.00217),  # prostate
             # no tissue
-            (TILES.NO_TISSUE, 6.677516254309149e-07),
-            (TILES.NO_TISSUE2, 7.505964521573081e-06),
-            (TILES.NO_TISSUE_LINE, 0.00028575083246431935),
-            (TILES.NO_TISSUE_RED_PEN, 0.2051245888881937),
-            (TILES.NO_TISSUE_GREEN_PEN, 0.28993597176882124),
+            (TILES.NO_TISSUE, 0.0),
+            (TILES.NO_TISSUE2, 0.0),
+            (TILES.NO_TISSUE_LINE, 0.00026),
+            (TILES.NO_TISSUE_RED_PEN, 0.20831),
+            (TILES.NO_TISSUE_GREEN_PEN, 0.23808),
         ),
     )
     def it_knows_nuclei_score(self, tile_img, expected_score):
@@ -62,86 +54,86 @@ def it_knows_nuclei_score(self, tile_img, expected_score):
         "tile_img, tissue, expected_score",
         (
             # level 0
-            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, True, 8.237973509211956e-05),
-            (TILES.LOW_NUCLEI_SCORE_LEVEL0, True, 0.029559623811739984),
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, True, 0.030890383076685523),
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, True, 0.5452530357003211),
+            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, True, 0.01494),
+            (TILES.LOW_NUCLEI_SCORE_LEVEL0, True, 0.13681),
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, True, 0.48335),
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, True, 0.01808),
             # level 1
             (
                 TILES.VERY_LOW_NUCLEI_SCORE_RED_PEN_LEVEL1,
                 True,
-                0.01610148106450298,
+                0.76041,
             ),  # breast - red pen
             (
                 TILES.LOW_NUCLEI_SCORE_LEVEL1,
                 True,
-                0.1887327689375471,
+                0.07103,
             ),  # breast - green pen
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, True, 0.04807442622295907),  # aorta
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, True, 0.39569),  # aorta
             (
                 TILES.MEDIUM_NUCLEI_SCORE_LEVEL1_2,
                 True,
-                0.23493428091089713,
+                0.15244,
             ),  # breast - green pen
             (
                 TILES.MEDIUM_NUCLEI_SCORE_GREEN_PEN_LEVEL1,
                 True,
-                0.6710679905294349,
+                0.15774,
             ),  # breast - green pen
             (
                 TILES.HIGH_NUCLEI_SCORE_RED_PEN_LEVEL1,
                 True,
-                0.30155274716806657,
+                0.21587,
             ),  # breast - red pen
             # level 2
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, True, 0.40466129363258146),  # prostate
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, True, 0.906540012633011),  # prostate
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, True, 0.02112),  # prostate
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, True, 0.18165),  # prostate
             # no tissue
-            (TILES.NO_TISSUE, True, 0.05299860529986053),
-            (TILES.NO_TISSUE2, True, 0.003337363966142684),
-            (TILES.NO_TISSUE_LINE, True, 0.3471366793342943),
-            (TILES.NO_TISSUE_RED_PEN, True, 0.7135526324697217),
-            (TILES.NO_TISSUE_GREEN_PEN, True, 0.6917688745017198),
-            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, False, 8.168825750149552e-05),
-            (TILES.LOW_NUCLEI_SCORE_LEVEL0, False, 0.024471282958984375),
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, False, 0.028293456864885436),
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, False, 0.5430199644432031),
+            (TILES.NO_TISSUE, True, 0.05021),
+            (TILES.NO_TISSUE2, True, 0.00203),
+            (TILES.NO_TISSUE_LINE, True, 0.31477),
+            (TILES.NO_TISSUE_RED_PEN, True, 0.72669),
+            (TILES.NO_TISSUE_GREEN_PEN, True, 0.56731),
+            (TILES.VERY_LOW_NUCLEI_SCORE_LEVEL0, False, 0.01482),
+            (TILES.LOW_NUCLEI_SCORE_LEVEL0, False, 0.11326),
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL0, False, 0.44272),
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL0, False, 0.01801),
             # level 1
             (
                 TILES.VERY_LOW_NUCLEI_SCORE_RED_PEN_LEVEL1,
                 False,
-                0.006328582763671875,
+                0.29887,
             ),  # breast - red pen
             (
                 TILES.LOW_NUCLEI_SCORE_LEVEL1,
                 False,
-                0.064971923828125,
+                0.02445,
             ),  # breast - green pen
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, False, 0.036724090576171875),
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL1, False, 0.30227),
             (
                 TILES.MEDIUM_NUCLEI_SCORE_LEVEL1_2,
                 False,
-                0.23455429077148438,
+                0.15219,
             ),  # breast - green pen
             (
                 TILES.MEDIUM_NUCLEI_SCORE_GREEN_PEN_LEVEL1,
                 False,
-                0.5416870117187,
+                0.12733,
             ),  # breast - green pen
             (
                 TILES.HIGH_NUCLEI_SCORE_RED_PEN_LEVEL1,
                 False,
-                0.2985572814941406,
+                0.21373,
             ),  # breast - red pen
             # level 2
-            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, False, 0.3309669494628906),  # prostate
-            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, False, 0.14234542846679688),  # prostate
+            (TILES.MEDIUM_NUCLEI_SCORE_LEVEL2, False, 0.01728),  # prostate
+            (TILES.HIGH_NUCLEI_SCORE_LEVEL2, False, 0.02852),  # prostate
             # no tissue
-            (TILES.NO_TISSUE, False, 0.0002899169921875),
-            (TILES.NO_TISSUE2, False, 0.000263214111328125),
-            (TILES.NO_TISSUE_LINE, False, 0.010105133056640625),
-            (TILES.NO_TISSUE_RED_PEN, False, 0.4020729064941406),
-            (TILES.NO_TISSUE_GREEN_PEN, False, 0.48259735107421875),
+            (TILES.NO_TISSUE, False, 0.00027),
+            (TILES.NO_TISSUE2, False, 0.00016),
+            (TILES.NO_TISSUE_LINE, False, 0.00916),
+            (TILES.NO_TISSUE_RED_PEN, False, 0.40948),
+            (TILES.NO_TISSUE_GREEN_PEN, False, 0.39577),
         ),
     )
     def it_knows_cellularity_score(self, tile_img, tissue, expected_score):