wip

tools/colorize_labels/colorize_labels.py

@@ -1,75 +1,39 @@
-"""
-Copyright (c) 2017-2024 Leonid Kostrykin, Biomedical Computer Vision Group, Heidelberg University
+import argparse
 
-Distributed under the MIT license.
-See file LICENSE for detail or copy at https://opensource.org/licenses/MIT
+import superdsm.io
+import superdsm.render
 
-Originally implemented in:
-https://github.com/BMCV/SuperDSM/blob/master/superdsm/render.py
-"""
 
-import numpy as np
+def _parse_hex_part(hex_part):
+    return int(f'0x{hex_part}')
 
 
-def shuffle_labels(labels, bg_label=None, seed=None):
-    """
-    Randomly shuffles the values of an integer-valued image.
+def color_hex_to_rgb_tuple(hex):
+    if hex.startswith('#'):
+        hex = hex[1:]
+    return (
+        _parse_hex_part(hex[0:2]),
+        _parse_hex_part(hex[2:4]),
+        _parse_hex_part(hex[4:6]),
+    )
 
-    :param labels:
-        An object of type ``numpy.ndarray`` corresponding to labeled segmentation masks.
 
-    :param bg_label:
-        If not ``None``, then this label stays fixed.
+if __name__ == '__main__':
 
-    :param seed:
-        The seed used for randomization.
+    parser = argparse.ArgumentParser()
+    parser.add_argument('input', type=str)
+    parser.add_argument('bg_label', type=int)
+    parser.add_argument('bg_color', type=str)
+    parser.add_argument('cmap', type=str, default='gist_rainbow')
+    parser.add_argument('seed', type=int)
+    args = parser.parse_args()
 
-    :return:
-        An object of type ``numpy.ndarray`` corresponding to ``labels`` with shuffled values (labels).
-    """
-    label_values0 = frozenset(labels.flatten())
-    if bg_label is not None: label_values0 -= {bg_label}
-    label_values0 = list(label_values0)
-    if seed is not None: np.random.seed(seed)
-    label_values1 = np.asarray(label_values0).copy()
-    np.random.shuffle(label_values1)
-    label_map = dict(zip(label_values0, label_values1))
-    result = np.zeros_like(labels)
-    for l in label_map.keys():
-        cc = (labels == l)
-        result[cc] = label_map[l]
-    return result
+    im = superdsm.io.imread(args.input)
 
-
-def colorize_labels(labels, bg_label=0, cmap='gist_rainbow', bg_color=(0,0,0), shuffle=None):
-    """
-    Returns a colorized representation of an integer-valued image.
-
-    :param labels:
-        An object of type ``numpy.ndarray`` corresponding to labeled segmentation masks.
-
-    :param bg_label:
-        Image areas with this label are forced to the color ``bg_color``.
-
-    :param cmap:
-        The colormap used to colorize the remaining labels (see `the list <https://matplotlib.org/stable/tutorials/colors/colormaps.html>`_).
-
-    :param bg_color:
-        The color used to represent the image regions with label ``bg_label`` (RGB).
-
-    :param shuffle:
-        If not ``None``, then used as ``seed`` to shuffle the labels before colorization, and not used otherwise.
-
-    :return:
-        An object of type ``numpy.ndarray`` corresponding to an RGB image.
-    """
-    if shuffle is not None:
-        labels = shuffle_labels(labels, bg_label=bg_label, seed=shuffle)
-    if isinstance(cmap, str):
-        cmap = plt.get_cmap(cmap)
-    img = cmap((labels - labels.min()) / float(labels.max() - labels.min()))
-    if img.shape[2] > 3: img = img[:,:,:3]
-    if bg_label is not None:
-        bg = (labels == bg_label)
-        img[bg] = np.asarray(bg_color)[None, None, :]
-    return img
+    superdsm.render.colorize_labels(
+        labels=im,
+        bg_label=args.bg_label,
+        cmap=args.cmap,
+        bg_color=color_hex_to_rgb_tuple(args.bg_color),
+        shuffle=args.seed,
+    )

tools/colorize_labels/colorize_labels.xml

@@ -0,0 +1,39 @@
+<tool id="colorize_labels" name="Colorize label map" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="20.05">
+    <description>with SuperDSM</description>
+    <macros>
+        <token name="@TOOL_VERSION@">0.2.0</token>
+        <token name="@VERSION_SUFFIX@">0</token>
+    </macros>
+    <edam_operations>
+        <edam_operation>operation_3443</edam_operation>
+    </edam_operations>
+    <xrefs>
+        <xref type="bio.tools">superdsm</xref>
+        <xref type="biii">superdsm</xref>
+    </xrefs>
+    <requirements> 
+        <requirement type="package" version="0.2.0">superdsm</requirement>
+    </requirements>
+    <inputs>
+        <param name="input" type="data" format="tiff,png" label="Input image (label map)" />
+        <param name="method_id" type="select" label="Thresholding method">
+            <option value="manual">Manual</option>
+            <option value="otsu" selected="True">Globally adaptive / Otsu</option>
+            <option value="li">Globally adaptive / Li's Minimum Cross Entropy</option>
+            <option value="isodata">Globally adaptive / Isodata</option>
+            <option value="yen">Globally adaptive / Yen</option>
+            <option value="loc_gaussian">Locally adaptive / Gaussian</option>
+            <option value="loc_median">Locally adaptive / Median</option>
+            <option value="loc_mean">Locally adaptive / Mean</option>
+        </param>
+    </inputs>
+    <outputs>
+       <data format="png" name="output" from_work_dir="out.png" />
+    </outputs>
+    <help>
+        Colorize label map for visualization.
+    </help>
+    <citations>
+        <citation type="doi">10.1109/TPAMI.2022.3185583</citation>
+    </citations>
+</tool>