ENH: Remove docstrings on itk.Image axis ordering. Remove ANTS option.

AGENTS.md

@@ -153,8 +153,6 @@ Version bumping: `bumpver update --patch`, `--minor`, or `--major`.
 - Masks are ITK images with integer labels. Keep anatomy group IDs consistent
   across segmenters.
 - Transforms are ITK composite transforms stored in compressed `.hdf` files.
-- State axis order and shape explicitly in every docstring and comment that
-  touches arrays.
 
 ## Implementation Role
 

CLAUDE.md

@@ -88,7 +88,6 @@ Regenerate it after any public API change: `py utils/generate_api_map.py`
   export by `vtk_to_usd.lps_points_to_usd` (USD +X=Left, +Y=Superior, +Z=Anterior)
 - Masks: ITK images with integer labels; consistent anatomy group IDs across all segmenters
 - Transforms: ITK composite transforms stored in `.hdf` files with compression
-- State axis order and shape explicitly in every docstring and comment that touches arrays
 
 ## Testing
 

docs/api/registration/time_series.rst

@@ -6,8 +6,7 @@ Time-Series Registration
 .. currentmodule:: physiomotion4d
 
 ``RegisterTimeSeriesImages`` registers ordered 3D image phases to a reference
-frame using ANTs, Greedy, ICON, or combined ``ANTS_ICON`` / ``greedy_ICON``
-methods.
+frame using Greedy, ICON, or combined ``Greedy_ICON`` methods.
 
 Class Reference
 ===============
@@ -29,7 +28,7 @@ Basic Usage
 
    images = [itk.imread(f"phase_{idx:02d}.mha") for idx in range(10)]
 
-   registrar = RegisterTimeSeriesImages(registration_method="ANTS")
+   registrar = RegisterTimeSeriesImages(registration_method="Greedy")
    registrar.set_fixed_image(images[0])
 
    result = registrar.register_time_series(

docs/api/workflows.rst

@@ -56,7 +56,7 @@ Convert Image to USD
        output_directory="./results",
        project_name="patient_001",
        segmentation_method="ChestTotalSegmentator",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
    final_usd = workflow.process()
@@ -157,7 +157,7 @@ High-Resolution 4D CT Reconstruction
        time_series_images=time_series_images,
        fixed_image=time_series_images[0],
        reference_frame=0,
-       registration_method="ANTS",
+       registration_method="Greedy_ICON",
    )
 
    result = workflow.run_workflow(upsample_to_fixed_resolution=True)

docs/cli_scripts/4dct_reconstruction.rst

@@ -47,8 +47,8 @@ Registration Options
    physiomotion4d-reconstruct-highres-4d-ct \
        --time-series-images frame_*.mha \
        --fixed-image highres_reference.mha \
-       --registration-method ANTS \
-       --ANTS-iterations 30 15 7 3 \
+       --registration-method Greedy_ICON \
+       --Greedy-iterations 30 15 7 3 \
        --prior-weight 0.5 \
        --output-dir ./results
 

docs/cli_scripts/best_practices.rst

@@ -171,7 +171,7 @@ Registration Failure
 **Solutions**:
    * Select better reference image (less motion, better quality)
    * Increase ``--registration-iterations``
-   * Try alternative registration method (``--registration-method ANTS``)
+   * Try alternative registration method (``--registration-method Greedy``)
    * Verify phases are temporally ordered correctly
 
 Memory Errors

docs/cli_scripts/heart_gated_ct.rst

@@ -112,7 +112,7 @@ Optional Arguments
      - Number of registration refinement iterations
    * - ``--registration-method``
      - ``ICON``
-     - Registration method: ``ICON`` or ``ANTS``
+     - Registration method: ``ICON`` or ``Greedy``
 
 Processing Pipeline
 ===================
@@ -215,14 +215,14 @@ Research Dataset Processing
            --project-name ${case_name}
    done
 
-With ANTs Registration
-----------------------
+With Greedy Registration
+------------------------
 
 .. code-block:: bash
 
    physiomotion4d-convert-image-to-usd cardiac.nrrd \
        --contrast \
-       --registration-method ANTS \
+       --registration-method Greedy \
        --registration-iterations 50
 
 Best Practices
@@ -289,7 +289,7 @@ Registration Failures
 **Registration not converging**
    * Try different reference image (better quality phase)
    * Increase ``--registration-iterations``
-   * Switch registration method (``--registration-method ANTS``)
+   * Switch registration method (``--registration-method Greedy``)
 
 **Excessive deformation**
    * Verify sufficient temporal overlap between phases

docs/developer/registration_images.rst

@@ -42,7 +42,7 @@ Time Series
 
    images = [itk.imread(f"phase_{idx:02d}.mha") for idx in range(10)]
 
-   registrar = RegisterTimeSeriesImages(registration_method="ANTS")
+   registrar = RegisterTimeSeriesImages(registration_method="Greedy")
    registrar.set_fixed_image(images[0])
    result = registrar.register_time_series(
        moving_images=images,

docs/developer/workflows.rst

@@ -39,7 +39,7 @@ Workflow Example
        contrast_enhanced=True,
        output_directory="./results",
        project_name="patient_001",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
    final_usd = workflow.process()

docs/examples.rst

@@ -35,7 +35,7 @@ Complete end-to-end cardiac CT processing:
        contrast_enhanced=True,
        output_directory="./results",
        project_name="patient_001",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
    # Run complete workflow
@@ -66,7 +66,7 @@ phase images:
        time_series_images=time_series_images,
        fixed_image=fixed_image,
        reference_frame=0,
-       registration_method="ANTS",
+       registration_method="Greedy_ICON",
    )
 
    result = workflow.run_workflow(upsample_to_fixed_resolution=True)
@@ -411,7 +411,7 @@ Batch process multiple datasets:
            contrast_enhanced=True,
            output_directory=f"results/{patient_id}",
            project_name=patient_id,
-           registration_method="ANTS",
+           registration_method="ICON",
        )
 
        try:
@@ -481,7 +481,7 @@ Run the supported end-to-end workflow API:
        contrast_enhanced=True,
        output_directory="./results",
        project_name="cardiac_model",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
    final_usd = workflow.process()

docs/quickstart.rst

@@ -72,7 +72,7 @@ CUDA-capable GPU are required for practical runtime.
    python -c "from physiomotion4d import DataDownloadTools; DataDownloadTools.DownloadSlicerHeartCTData('data/test')"
 
    physiomotion4d-convert-image-to-usd data/test/TruncalValve_4DCT.seq.nrrd \
-       --registration-method ANTS \
+       --registration-method ICON \
        --output-dir output/quickstart \
        --project-name slicer_heart_quickstart
 
@@ -116,7 +116,7 @@ For more control, use the Python API:
        contrast_enhanced=True,
        output_directory="./results",
        project_name="cardiac_model",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
 **Step 3: Run the workflow**
@@ -151,7 +151,7 @@ For more control over individual steps:
        contrast_enhanced=True,
        output_directory="./results",
        project_name="cardiac_model",
-       registration_method="ANTS",
+       registration_method="ICON",
    )
 
    final_usd = workflow.process()
@@ -314,7 +314,7 @@ Common Issues
 
 * Resample or crop the input image before running the workflow
 * Process fewer frames at once
-* Use ANTs registration with ``--registration-method ANTS`` when CUDA is unavailable
+* Use Greedy registration with ``--registration-method Greedy`` when CUDA is unavailable
 
 **Segmentation quality issues**
 

docs/troubleshooting.rst

@@ -15,7 +15,7 @@ CUDA Out of Memory
 **Solutions**:
 
 1. Resample or crop the input image before running the workflow.
-2. Use ``--registration-method ANTS`` when CUDA is unavailable.
+2. Use ``--registration-method Greedy`` when CUDA is unavailable.
 3. Process fewer frames per run.
 
 CUDA Version Mismatch
@@ -93,7 +93,7 @@ Registration Not Converging
 
    .. code-block:: bash
 
-      physiomotion4d-convert-image-to-usd cardiac_4d.nrrd --registration-method ANTS
+      physiomotion4d-convert-image-to-usd cardiac_4d.nrrd --registration-method Greedy
 
 3. Check image orientation and spacing
 

docs/tutorials.rst

@@ -244,7 +244,7 @@ Script
    ``tutorials/tutorial_08_dirlab_pca_time_series.py``
 
 Workflow
-   ``RegisterTimeSeriesImages`` with ``registration_method='ANTS_ICON'`` and
+   ``RegisterTimeSeriesImages`` with ``registration_method='Greedy_ICON'`` and
    ``TransformTools``
 
 Dataset

experiments/Reconstruct4DCT/reconstruct_4d_ct_class.py

@@ -53,22 +53,22 @@
     num_files = len(files)
     reference_image_num = num_files // 2
 
-    # Registration parameters - only ANTs for quick run
-    registration_methods = ["ANTS", "ICON", "ANTS_ICON"]
-    number_of_iterations_list = [[8, 4, 1], 5, [[8, 4, 1], 5]]  # For ANTs and ICON
+    # Registration parameters - only Greedy for quick run
+    registration_methods = ["Greedy", "ICON", "Greedy_ICON"]
+    number_of_iterations_list = [[8, 4, 1], 5, [[8, 4, 1], 5]]  # For Greedy and ICON
 else:
     print("=== FULL RUN MODE ===")
     num_files = len(files)
     files_indx = list(range(num_files))
     reference_image_num = 7
 
-    # Registration parameters - both ANTs and ICON for full run
-    registration_methods = ["ANTS"]  # , "ICON", "ANTS_ICON"]
+    # Registration parameters - Greedy and ICON for full run
+    registration_methods = ["Greedy"]  # , "ICON", "Greedy_ICON"]
     number_of_iterations_list = [
         [30, 15, 7, 3],
-    ]  # For ANTs
+    ]  # For Greedy
     # 20,  # For ICON
-    # [[30, 15, 7, 3], 20],  # For ANTS_ICON
+    # [[30, 15, 7, 3], 20],  # For Greedy_ICON
     # ]
 
 # Common parameters
@@ -144,12 +144,12 @@
     registrar.set_fixed_image(fixed_image)
 
     # Set iterations based on registration method
-    if registration_method == "ANTS":
-        registrar.set_number_of_iterations_ANTS(number_of_iterations)
+    if registration_method == "Greedy":
+        registrar.set_number_of_iterations_greedy(number_of_iterations)
     elif registration_method == "ICON":
         registrar.set_number_of_iterations_ICON(number_of_iterations)
-    elif registration_method == "ANTS_ICON":
-        registrar.set_number_of_iterations_ANTS(number_of_iterations[0])
+    elif registration_method == "Greedy_ICON":
+        registrar.set_number_of_iterations_greedy(number_of_iterations[0])
         registrar.set_number_of_iterations_ICON(number_of_iterations[1])
 
     # Perform registration

src/physiomotion4d/cli/convert_image_to_usd.py

@@ -87,9 +87,9 @@ def main() -> int:
     )
     parser.add_argument(
         "--registration-method",
-        choices=["ANTS", "ICON"],
+        choices=["Greedy", "ICON"],
         default="ICON",
-        help="Registration method to use: ANTS or ICON (default: ICON)",
+        help="Registration method to use: Greedy or ICON (default: ICON)",
     )
     parser.add_argument(
         "--fps",

src/physiomotion4d/contour_tools.py

@@ -214,9 +214,7 @@ def create_mask_from_mesh(
             voxel_grid_origin_lps[2] + voxel_pitch * binary_array.shape[2]
         )
 
-        # Create ITK image from the voxel array
-        # ITK uses ZYX ordering (numpy array convention), trimesh uses XYZ
-        # Need to transpose: (X, Y, Z) -> (Z, Y, X)
+        # transpose to match trimesh XYZ convention
         binary_array_zyx = np.transpose(binary_array, (2, 1, 0))
         binary_array_flip = np.flip(binary_array_zyx, axis=0)
         binary_image = itk.GetImageFromArray(binary_array_flip)

src/physiomotion4d/convert_image_4d_to_3d.py

@@ -32,7 +32,7 @@
 
 
 class ConvertImage4DTo3D(PhysioMotion4DBase):
-    """Split a 3D/4D ITK image (X, Y, Z [, T]) into a list of 3D ITK images."""
+    """Split a 3D/4D ITK image into a list of 3D ITK images."""
 
     def __init__(self, log_level: int | str = logging.INFO) -> None:
         """Initialize the 4D-to-3D image converter.
@@ -178,7 +178,7 @@ def _build_frames(
         spacing_3d: np.ndarray,
         direction_3d: np.ndarray,
     ) -> None:
-        """Materialize ``self.img_3d`` from a (T, Z, Y, X) array + geometry."""
+        """Materialize ``self.img_3d`` from a time-series array + geometry."""
         direction_matrix = itk.matrix_from_array(np.ascontiguousarray(direction_3d))
         self.img_3d = []
         for t in range(arr_4d.shape[0]):

src/physiomotion4d/image_tools.py

@@ -104,17 +104,11 @@ def convert_itk_image_to_sitk(self, itk_image: itk.Image) -> sitk.Image:
         Returns:
             SimpleITK image with identical data and metadata
 
-        Note:
-            Memory layout is preserved during conversion. Both ITK and SimpleITK
-            use (z, y, x) ordering for 3D images in numpy arrays.
-
         Example:
             >>> tools = ImageTools()
             >>> itk_image = itk.imread('image.nii.gz')
             >>> sitk_image = tools.convert_itk_image_to_sitk(itk_image)
         """
-        # Get numpy array from ITK image
-        # ITK array is in (z, y, x) or (z, y, x, components) format
         array = itk.array_from_image(itk_image)
 
         # Get image metadata
@@ -129,11 +123,8 @@ def convert_itk_image_to_sitk(self, itk_image: itk.Image) -> sitk.Image:
             is_vector = n_components > 1
 
         if is_vector:
-            # For vector images, SimpleITK expects (z, y, x, components) which is what we have
-            # Create SimpleITK image from numpy array
             sitk_image = sitk.GetImageFromArray(array, isVector=True)
         else:
-            # For scalar images, array is (z, y, x)
             sitk_image = sitk.GetImageFromArray(array, isVector=False)
 
         # Set metadata
@@ -162,17 +153,11 @@ def convert_sitk_image_to_itk(self, sitk_image: sitk.Image) -> itk.Image:
         Returns:
             ITK image with identical data and metadata
 
-        Note:
-            Memory layout is preserved during conversion. Both SimpleITK and ITK
-            use (z, y, x) ordering for 3D images in numpy arrays.
-
         Example:
             >>> tools = ImageTools()
             >>> sitk_image = sitk.ReadImage('image.nii.gz')
             >>> itk_image = tools.convert_sitk_image_to_itk(sitk_image)
         """
-        # Get numpy array from SimpleITK image
-        # SimpleITK array is in (z, y, x) or (z, y, x, components) format
         array = sitk.GetArrayFromImage(sitk_image)
 
         # Get image metadata

src/physiomotion4d/labelmap_tools.py

@@ -61,11 +61,6 @@ def convert_labelmap_to_mask(
         even on anisotropic grids; each per-axis count is clamped to at least
         1 voxel when ``dilation_in_mm > 0``.
 
-        Axis ordering: the labelmap is a scalar 3D ``itk.Image`` in ITK
-        world-axis order (X, Y, Z). All thresholding is performed on the numpy
-        view (Z, Y, X) and written back through ``CopyInformation``, so origin,
-        spacing, and direction are preserved.
-
         Args:
             labelmap: Multi-label or binary ``itk.Image``. Any non-zero voxel
                 that is not excluded is treated as foreground.
@@ -132,11 +127,6 @@ def create_distance_map(
         this continuous encoding gives every region a smoothly varying signal
         while preserving label identity.
 
-        Axis ordering: ``labelmap`` is a scalar 3D ``itk.Image`` in ITK
-        world-axis order (X, Y, Z). All work is done on the numpy view
-        (Z, Y, X) and written back through ``CopyInformation``, so origin,
-        spacing, and direction are preserved.
-
         Args:
             labelmap: Multi-label (or binary) ``itk.Image`` of integer labels.
             max_distance_mm: Distance clip, in millimeters. Default 20.0.

src/physiomotion4d/register_images_ants.py

@@ -193,7 +193,6 @@ def _itk_to_ants_image(
             raise ValueError(f"Unsupported dtype: {dtype}")
 
         if is_vector:
-            # Vector images: ITK gives (z,y,x,components) or (y,x,components)
             spatial_shape = data.shape[:-1]  # drop components
         else:
             spatial_shape = data.shape