Support multiple regions (#26)

* points to brainrender * update testing * refactor * refactor * badges * notes * initial inj finder commit (#11) * formatting with black * add .DS_store to gitignore * typo * update imports * update imports * deal with no obj path given * error typo * move utils too imlib * remove amap imports * optional logging to file * tidy * registration moved to imlib * keep largest object * reduce RAM usage * remove unnecessary logging args * adds lesion and track estimation and generic tools * add missing dir support for cell export * add missing dir support for cell export * reduce mem requirements * remove old code * remove old code * refactor scaling * fix failing heatmap test * update docs * ensure output directory exists * make pixel sizes mandatory * make pixel sizes mandatory (#17) * move functionality to imlib * remove hard coded smoothing * Refactor (#19) * make pixel sizes mandatory * move functionality to imlib * remove hard coded smoothing * update docs * Injectionsite (#20) * refactors code from SL to import from neuro and imlib only * moves amap_vis to neuro and removes amap and cellfinder dependencies; * format with black * amap_vis * bump * remove cellfinder dependency * cli entry point for fibre track segmentation * refactor seg tools * start manual seg * remove hardocoding in manual region segmentation * refactor paths * bump * update docs * update docs * update docs * update docs * support multiple objects * color objects * bump * update reqs Co-authored-by: Federico Claudi <federicoclaudi@protonmail.com> Co-authored-by: stephen <stephen.lenzi@gmail.com>
SainsburyWellcomeCentre · Apr 20, 2020 · 2e02dd2 · 2e02dd2
1 parent c61036e
commit 2e02dd2
Show file tree

Hide file tree

Showing 3 changed files with 71 additions and 11 deletions.
diff --git a/neuro/brain_render_tools.py b/neuro/brain_render_tools.py
@@ -1,7 +1,11 @@
+import numpy as np
+from pathlib import Path
 from brainrender.Utils.image import reorient_image, marching_cubes_to_obj
 from brainrender.scene import Scene
 from skimage import measure
 
+from imlib.general.pathlib import append_to_pathlib_stem
+
 from neuro.atlas_tools.custom_atlas_structures import (
     get_arbitrary_structure_mask_from_custom_atlas,
 )
@@ -44,21 +48,55 @@ def volume_to_vector_array_to_obj_file(
     voxel_size=10,
     orientation="coronal",
     step_size=1,
+    threshold=0,
+    deal_with_regions_separately=False,
 ):
 
     oriented_binary = reorient_image(
         image, invert_axes=invert_axes, orientation=orientation
     )
-
+    #
+    if deal_with_regions_separately:
+        for label_id in np.unique(oriented_binary):
+            filename = append_to_pathlib_stem(
+                Path(output_path), "_" + str(label_id)
+            )
+            image = oriented_binary == label_id
+            extract_and_save_object(
+                image,
+                filename,
+                voxel_size=voxel_size,
+                threshold=threshold,
+                step_size=step_size,
+            )
+    else:
+        extract_and_save_object(
+            oriented_binary,
+            output_path,
+            voxel_size=voxel_size,
+            threshold=threshold,
+            step_size=step_size,
+        )
+
+
+def extract_and_save_object(
+    image, output_file_name, voxel_size=10, threshold=0, step_size=1
+):
     verts, faces, normals, values = measure.marching_cubes_lewiner(
-        oriented_binary, 0, step_size=step_size
+        image, threshold, step_size=step_size
     )
+    verts, faces = convert_obj_to_br(verts, faces, voxel_size=voxel_size)
+    marching_cubes_to_obj(
+        (verts, faces, normals, values), str(output_file_name)
+    )
+
 
+def convert_obj_to_br(verts, faces, voxel_size=10):
     if voxel_size is not 1:
         verts = verts * voxel_size
 
     faces = faces + 1
-    marching_cubes_to_obj((verts, faces, normals, values), str(output_path))
+    return verts, faces
 
 
 def visualize_obj(obj_path, *args, color="lightcoral", **kwargs):

diff --git a/neuro/segmentation/manual_region_segmentation/segment.py b/neuro/segmentation/manual_region_segmentation/segment.py
@@ -3,10 +3,13 @@
 
 from pathlib import Path
 import numpy as np
+from glob import glob
 
 from PySide2.QtWidgets import QApplication
 from brainrender.scene import Scene
-from imlib.general.system import delete_temp
+from imlib.general.system import delete_temp, ensure_directory_exists
+from imlib.plotting.colors import get_random_vtkplotter_color
+
 
 from neuro.generic_neuro_tools import (
     transform_image_to_standard_space,
@@ -27,8 +30,13 @@ def __init__(self, registration_output_folder, downsampled_image):
         self.registration_output_folder = Path(registration_output_folder)
         self.downsampled_image = self.join(downsampled_image)
 
-        self.regions_object_file = self.join("regions.obj")
-        self.regions_image_file = self.join("regions.nii")
+        self.regions_directory = self.join("segmented_regions")
+
+        self.regions_object_file_basename = (
+            self.regions_directory / "region.obj"
+        )
+
+        self.regions_image_file = self.regions_directory / "regions.nii"
 
         self.tmp__inverse_transformed_image = self.join(
             "image_standard_space.nii"
@@ -40,9 +48,14 @@ def __init__(self, registration_output_folder, downsampled_image):
             "inverse_transform_error.txt"
         )
 
+        self.prep()
+
     def join(self, filename):
         return self.registration_output_folder / filename
 
+    def prep(self):
+        ensure_directory_exists(self.regions_directory)
+
 
 def run(
     image,
@@ -88,11 +101,15 @@ def run(
 
         @viewer.bind_key("Control-S")
         def add_region(viewer):
-            print(f"\nSaving regions to: {paths.regions_object_file}")
+            print(f"\nSaving regions to: {paths.regions_directory}")
             # return image back to original orientation (reoriented for napari)
             data = np.swapaxes(labels_layer.data, 2, 0)
 
-            volume_to_vector_array_to_obj_file(data, paths.regions_object_file)
+            volume_to_vector_array_to_obj_file(
+                data,
+                paths.regions_object_file_basename,
+                deal_with_regions_separately=True,
+            )
             if save_segmented_image:
                 save_brain(
                     data, paths.downsampled_image, paths.regions_image_file,
@@ -106,9 +123,14 @@ def add_region(viewer):
         delete_temp(paths.registration_output_folder, paths)
 
     if preview:
+
         print("\nPreviewing in brainrender")
         scene = Scene()
-        scene.add_from_file(str(paths.regions_object_file), c="coral")
+        obj_files = glob(str(paths.regions_directory) + "/*.obj")
+        for obj_file in obj_files:
+            act = scene.add_from_file(
+                obj_file, c=get_random_vtkplotter_color(), alpha=0.8
+            )
         scene.render()
 
 

diff --git a/setup.py b/setup.py
@@ -6,14 +6,14 @@
     "pandas<=0.25.3,>=0.25.1",
     "napari",
     "brainrender",
-    "imlib",
+    "imlib >= 0.0.18",
     "brainio",
 ]
 
 
 setup(
     name="neuro",
-    version="0.0.8rc0",
+    version="0.0.8rc1",
     description="Visualisation and analysis of brain imaging data",
     install_requires=requirements,
     extras_require={