def run(input_directories: list,
output_path: str,
segment_type: list,
open_viewer=True):
flair_input_path = input_directories[0]
t1_input_path = input_directories[1]
ir_input_path = input_directories[2]
logging.info("Starting brain pipeline")
flair_array = read_input_path_as_np_array(flair_input_path)
t1_array = read_input_path_as_np_array(t1_input_path)
ir_array = read_input_path_as_np_array(ir_input_path)
segmented_array = brain_model.predict(flair_array, t1_array, ir_array)
meshes = [
generate_mesh(segment, 0)
for segment in seperate_segmentation(segmented_array,
unique_values=segment_type)
]
if open_viewer:
meshes = convert_meshes_trimesh(meshes)
segment_dict = get_seg_types(this_plid)
mesh_names = [k for k, v in segment_dict.items() if v in segment_type]
view_mesh(
meshes=meshes,
mesh_names=mesh_names,
output_file=output_path,
plid=this_plid,
)
else:
write_mesh_as_glb_with_colour(meshes, output_path)
logging.info("Brain pipeline finished successfully")
def run(input_path: str,
output_path: str,
segment_type: list,
open_viewer=True) -> None:
logging.info("Starting abdominal pipeline")
image_array = read_input_path_as_np_array(input_path)
image_array = downscale_and_conditionally_crop(image_array)
# NOTE: Numpy array is flipped in the Y axis here as this is the specific image input for the NiftyNet model
image_array = flip_numpy_array_dimensions_y_only(image_array)
nifti_image = convert_np_ndarray_to_nifti_image(image_array)
initial_nifti_output_file_path = abdominal_model.get_input_path()
write_nifti_image(nifti_image, initial_nifti_output_file_path)
segmented_nifti_output_file_path = abdominal_model.predict()
segmented_array = read_nifti_as_np_array(segmented_nifti_output_file_path,
normalise=False)
meshes = [
generate_mesh(segment, 0)
for segment in seperate_segmentation(segmented_array,
unique_values=segment_type)
]
if open_viewer:
metadata = get_metadata(input_path)
meshes = convert_meshes_trimesh(meshes)
segment_dict = get_seg_types(this_plid)
mesh_names = [k for k, v in segment_dict.items() if v in segment_type]
view_mesh(
meshes=meshes,
mesh_names=mesh_names,
output_file=output_path,
patient_data=metadata,
plid=this_plid,
scan_path=input_path,
)
else:
write_mesh_as_glb_with_colour(meshes, output_path)
abdominal_model.cleanup()
logging.info("Abdominal pipeline finished successfully")
def run(
input_path: str, output_path: str, segment_type: list, open_viewer=True
) -> None:
logging.info("Starting kidney pipeline")
image = read_nifti_image(input_path)
initial_nifti_output_file_path = kidney_model.get_input_path()
write_nifti_image(image, initial_nifti_output_file_path)
segmented_nifti_output_file_path = kidney_model.predict()
segmented_array = read_nifti_as_np_array(
segmented_nifti_output_file_path, normalise=False
)
meshes = [
generate_mesh(segment, 0)
for segment in seperate_segmentation(
segmented_array, unique_values=segment_type
)
]
if open_viewer:
metadata = get_metadata(input_path)
meshes = convert_meshes_trimesh(meshes)
segment_dict = get_seg_types(this_plid)
mesh_names = [k for k, v in segment_dict.items() if v in segment_type]
view_mesh(
meshes=meshes,
mesh_names=mesh_names,
output_file=output_path,
patient_data=metadata,
plid=this_plid,
scan_path=input_path,
)
else:
write_mesh_as_glb_with_colour(meshes, output_path)
kidney_model.cleanup()
logging.info("Kidney pipeline finished successfully")
def run(input_path: str,
output_path: str,
segment_type: list,
open_viewer=True) -> None:
logging.info("Starting bone pipeline")
image_data = read_input_path_as_np_array(input_path)
image_data = downscale_and_conditionally_crop(image_data)
meshes = [generate_mesh(image_data, bone_hu_threshold)]
if open_viewer:
metadata = get_metadata(input_path)
meshes = convert_meshes_trimesh(meshes)
segment_dict = get_seg_types(this_plid)
mesh_names = [k for k, v in segment_dict.items() if v in segment_type]
view_mesh(
meshes=meshes,
mesh_names=mesh_names,
output_file=output_path,
patient_data=metadata,
plid=this_plid,
scan_path=input_path,
)
else:
write_mesh_as_glb_with_colour(meshes, output_path)
logging.info("Bone pipeline finished successfully")
def run(input_path: str,
output_path: str,
segment_type: list,
open_viewer=True) -> None:
logging.info("Starting lung pipeline")
image_data = read_input_path_as_np_array(input_path)
image_data = downscale_and_conditionally_crop(image_data)
segmented_lung, segmented_airway = perform_lung_segmentation(image_data)
meshes = []
if 1 in segment_type: # lung segmentation
meshes.append(generate_mesh(segmented_lung, hu_threshold))
if 2 in segment_type: # airway segmentation
meshes.append(generate_mesh(segmented_airway, hu_threshold))
if len(meshes) == 0:
raise Exception(
"No valid segmentation specified, segmentation type must be either 1 or 2"
)
if open_viewer:
metadata = get_metadata(input_path)
meshes = convert_meshes_trimesh(meshes)
segment_dict = get_seg_types(this_plid)
mesh_names = [k for k, v in segment_dict.items() if v in segment_type]
view_mesh(
meshes=meshes,
mesh_names=mesh_names,
output_file=output_path,
patient_data=metadata,
plid=this_plid,
scan_path=input_path,
)
else:
write_mesh_as_glb_with_colour(meshes, output_path)
logging.info("Lung pipeline finished successfully")