Dear Slicer community,
i switched to 3D Slicer 5.6.0 and so reinstalled the python modules that i use, for exemple openslide.
But after “coding a bit” (ill post at the end of this post what i mean by that), i restarded 3D slicer and was suprised to see this already in my python console with a loaded volume :
Level: 0, Width: 105728, Height: 77952, Downsample Factor: 1.0, Magnification: 40.0x, Resolution: 0.000442 mm/pixel
Level: 1, Width: 52864, Height: 38976, Downsample Factor: 2.0, Magnification: 20.0x, Resolution: 0.000884 mm/pixel
Level: 2, Width: 26432, Height: 19488, Downsample Factor: 4.0, Magnification: 10.0x, Resolution: 0.001768 mm/pixel
Level: 3, Width: 13216, Height: 9744, Downsample Factor: 8.0, Magnification: 5.0x, Resolution: 0.003536 mm/pixel
Level: 4, Width: 6608, Height: 4872, Downsample Factor: 16.0, Magnification: 2.5x, Resolution: 0.007072 mm/pixel
Level: 5, Width: 3304, Height: 2436, Downsample Factor: 32.0, Magnification: 1.25x, Resolution: 0.014144 mm/pixel
Level: 6, Width: 1652, Height: 1218, Downsample Factor: 64.0, Magnification: 0.625x, Resolution: 0.028288 mm/pixel
Level: 7, Width: 826, Height: 609, Downsample Factor: 128.0, Magnification: 0.3125x, Resolution: 0.056576 mm/pixel
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/home/nicolas.tempier/Slicer-5.6.0-linux-amd64/lib/Python/lib/python3.9/imp.py", line 169, in load_source
module = _exec(spec, sys.modules[name])
File "<frozen importlib._bootstrap>", line 613, in _exec
File "<frozen importlib._bootstrap_external>", line 850, in exec_module
File "<frozen importlib._bootstrap>", line 228, in _call_with_frames_removed
File "/home/nicolas.tempier/slicer_modules/H3H/Cryoslice_loader/openslide_slicer.py", line 176, in <module>
image_fullsize_level6.load_image()
File "/home/nicolas.tempier/slicer_modules/H3H/Cryoslice_loader/openslide_slicer.py", line 66, in load_image
voxel_spacing = tuple(i * return_vowel_spacing_from_level_in_mm(self.image_path, self.level) for i in base_voxel_spacing)
File "/home/nicolas.tempier/slicer_modules/H3H/Cryoslice_loader/openslide_slicer.py", line 66, in <genexpr>
voxel_spacing = tuple(i * return_vowel_spacing_from_level_in_mm(self.image_path, self.level) for i in base_voxel_spacing)
NameError: name 'return_vowel_spacing_from_level_in_mm' is not defined
[Qt] libpng warning: iCCP: profile 'ICC Profile': 'CMYK': invalid ICC profile color space
[Qt] libpng warning: iCCP: known incorrect sRGB profile
[Qt] loadSourceAsModule - Failed to load file "/home/nicolas.tempier/slicer_modules/H3H/Cryoslice_loader/openslide_slicer.py" as module "openslide_slicer" !
[Qt] Fail to instantiate module "openslide_slicer"
[Qt] The following modules failed to be instantiated:
[Qt] openslide_slicer
Loading Slicer RC file [/home/nicolas.tempier/.slicerrc.py]
how do i get rid of this without having to reinstall 3D slicer everytime i restart it please ?
PS : to install openslide i “had” to mannually copy paste
“openslide” and “openslide_python-1.2.0.dist-info” to sites-packages and that could be a reason but I never managed to install it otherwise.
thank you in advance for your help,
best,
Nicolas
PS 2 :
here is what i executed before restarting :
import slicer
import numpy as np
import vtk
from vtk.util.vtkImageImportFromArray import vtkImageImportFromArray
from vtk.util import numpy_support
from PIL import Image
import timeit
import gc
import openslide
class ImageLoader:
def __init__(self, image_path, level, size=None, somewhere_coordinates=(0, 0), volume_name='image_loaded'):
self.image_path = image_path
self.level = level
self.size = size
self.somewhere_coordinates = somewhere_coordinates
self.volume_name = volume_name
def load_image(self):
# Charger l'image NDPI
self.image_ndpi = openslide.OpenSlide(self.image_path)
# Obtenir le facteur de sous-échantillonnage du niveau
self.downsample_factor = self.image_ndpi.level_downsamples[self.level]
# Convertir les coordonnées du niveau cible en coordonnées du niveau 0
somewhere_coordinates_level_0 = (int(self.somewhere_coordinates[1] * self.downsample_factor),
int(self.somewhere_coordinates[0] * self.downsample_factor))
# Sélectionner résolution
# print(f"Level selectionné : {self.level}")
if self.size is None:
self.size = self.image_ndpi.level_dimensions[self.level]
# Lire l'image en utilisant les coordonnées du niveau 0
self.image_matrix = self.image_ndpi.read_region(somewhere_coordinates_level_0, self.level, self.size)
# Convertir l'image en un tableau NumPy et créer un objet VTK Image
self.image_array = np.array(self.image_matrix)[:, :, :3]
self.image_array = np.swapaxes(self.image_array, 0, 1)
self.image_array = np.ascontiguousarray(self.image_array, dtype=np.uint8)
# Créer une image VTK avec 3 dimensions spatiales et 3 dimensions d'intensité pour les canaux RGB
self.vtk_image_rgb = vtk.vtkImageData()
self.vtk_image_rgb.SetDimensions(self.image_array.shape[1], self.image_array.shape[0], 1)
self.vtk_image_rgb.AllocateScalars(vtk.VTK_UNSIGNED_CHAR, 3)
self.vtk_image_rgb.SetSpacing(1, 1, 1)
# Convertir le tableau NumPy en une liste d'octets
self.image_list = self.image_array.flatten()
# Convertir la liste d'octets en un tableau VTK
self.vtk_array = numpy_support.numpy_to_vtk(num_array=self.image_list, deep=True, array_type=vtk.VTK_UNSIGNED_CHAR)
# Associer le tableau VTK à un vtkUnsignedCharArray
self.rgb_array_vtk = vtk.vtkUnsignedCharArray()
self.rgb_array_vtk.SetNumberOfComponents(3)
self.rgb_array_vtk.SetArray(self.vtk_array, len(self.image_list), 1)
# Associer le tableau vtkUnsignedCharArray aux scalaires de l'image VTK
self.vtk_image_rgb.GetPointData().SetScalars(self.rgb_array_vtk)
# Ajouter l'image à un nœud de volume dans 3D Slicer
self.volume_node = slicer.mrmlScene.AddNewNodeByClass('vtkMRMLVectorVolumeNode', self.volume_name)
base_voxel_spacing = (1, 1, 1)
voxel_spacing = tuple(i * return_vowel_spacing_from_level_in_mm(self.image_path, self.level) for i in base_voxel_spacing)
self.volume_node.SetSpacing(voxel_spacing)
self.volume_node.SetAndObserveImageData(self.vtk_image_rgb)
slicer.app.applicationLogic().PropagateVolumeSelection(0)
# remettre le volume au bon endroit si somewhere_coordinates n'est pas (0,0)
if self.somewhere_coordinates != (0,0):
self.volume_node.SetOrigin(self.somewhere_coordinates[0]*voxel_spacing[0],self.somewhere_coordinates[1]*voxel_spacing[0],0)
def remove_node(self):
node = slicer.mrmlScene.GetFirstNodeByName(self.volume_name)
if node is not None:
slicer.mrmlScene.RemoveNode(node)
def return_vowel_spacing_from_level_in_mm(image_path, level):
# Charger l'image NDPI
slide = openslide.OpenSlide(image_path)
# Suppose que le grossissement au niveau 0 est de 40x
magnification_at_level_0 = 40
# Résolution au niveau 0 en nm/pixel
resolution_at_level_0_nm_per_pixel = 442
resolution_at_target_level_mm_per_pixel = resolution_at_level_0_nm_per_pixel/1e6 * slide.level_downsamples[level]
return resolution_at_target_level_mm_per_pixel
def clear_memory(self):
# Libérer la mémoire des objets
# del self.image_ndpi
# del self.image_matrix
# del self.image_array
# del self.image_list
del self.vtk_array
del self.rgb_array_vtk
gc.collect()
def display_downsample_factors(image_path):
# Charger l'image NDPI
slide = openslide.OpenSlide(image_path)
# Suppose que le grossissement au niveau 0 est de 40x
magnification_at_level_0 = 40
# Résolution au niveau 0 en nm/pixel
resolution_at_level_0_nm_per_pixel = 442
for i in range(slide.level_count):
width, height = slide.level_dimensions[i]
downsample = slide.level_downsamples[i]
magnification = magnification_at_level_0 / downsample
# Convertir la résolution de nm/pixel à mm/pixel
resolution_at_level_mm_per_pixel = (resolution_at_level_0_nm_per_pixel * downsample) / 1e6
print(f'Level: {i}, Width: {width}, Height: {height}, Downsample Factor: {downsample}, Magnification: {magnification}x, Resolution: {resolution_at_level_mm_per_pixel} mm/pixel')
def set_voxel_spacing(volume_name, voxel_spacing):
volume_node = slicer.util.getNode(volume_name)
if volume_node is not None:
volume_node.SetSpacing(voxel_spacing)
def display_image_levels_and_sizes(image_path):
# Charger l'image NDPI
image_ndpi = openslide.OpenSlide(image_path)
# Obtenir le nombre de niveaux de l'image
num_levels = len(image_ndpi.level_dimensions)
# Afficher les niveaux de grossissement et les tailles d'image correspondantes
print("Niveau de grossissement et tailles d'image correspondantes :")
for i in range(num_levels):
level_dimensions = image_ndpi.level_dimensions[i]
downsample_factor = image_ndpi.level_downsamples[i]
print(f"Niveau {i}: facteur de sous-échantillonnage = {downsample_factor}, dimensions = {level_dimensions}")
def ask_for_coordinates():
user_input = input("Entrez les coordonnées x et y du premier point, séparées par une virgule : ")
x1, y1 = map(int, user_input.split(','))
user_input = input("Entrez les coordonnées x et y du deuxième point, séparées par une virgule : ")
x2, y2 = map(int, user_input.split(','))
return (x1, y1), (x2, y2)
def get_corresponding_coords(image_path, low_res_level, high_res_level, x, y):
slide = openslide.OpenSlide(image_path)
# Récupère le facteur de sous-échantillonnage pour chaque niveau
low_res_downsample = slide.level_downsamples[low_res_level]
high_res_downsample = slide.level_downsamples[high_res_level]
# Convertit les coordonnées au niveau de résolution basse en coordonnées au niveau 0
x_at_level_0 = x * low_res_downsample
y_at_level_0 = y * low_res_downsample
# Convertit les coordonnées au niveau 0 en coordonnées au niveau de résolution haute
x_at_high_res_level = x_at_level_0 / high_res_downsample
y_at_high_res_level = y_at_level_0 / high_res_downsample
return x_at_high_res_level, y_at_high_res_level
# setup
image_path = "/network/lustre/iss02/lau-karachi/data_raw/Human/HYPOTHALAMUS-SANO/Homme/H3h/tests_ChAT/ChAT_Tests/NicolasTempier/H3h 440 ChAT_Test - 2023-11-28 20.28.02.ndpi"
display_downsample_factors(image_path)
# image_fullsize_level7.remove_node() # remove the node when done
# on veut charger l'image au level 6 taille entiere
image_fullsize_level6 = ImageLoader(image_path, 6, somewhere_coordinates=(0, 0), volume_name='image_fullsize_level6')
image_fullsize_level6.load_image()