Merge pull request #220 from DiamondLightSource/tomobarpadding

padding parameter for the horizontal detector (tomobar v. 2025.08)

Author: dkazanc

httomolibgpu/recon/algorithm.py

@@ -59,6 +59,7 @@ def FBP2d_astra(
     data: np.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     filter_type: str = "ram-lak",
     filter_parameter: Optional[float] = None,
     filter_d: Optional[float] = None,
@@ -81,6 +82,8 @@ def FBP2d_astra(
         An array of angles given in radians.
     center : float, optional
         The center of rotation (CoR).
+    detector_pad : int
+        Detector width padding with edge values to remove circle/arc type artifacts in the reconstruction.
     filter_type: str
         Type of projection filter, see ASTRA's API for all available options for filters.
     filter_parameter: float, optional
@@ -112,7 +115,7 @@ def FBP2d_astra(
         recon_size = data_shape[2]
 
     RecTools = _instantiate_direct_recon2d_class(
-        data, angles, center, recon_size, gpu_id
+        data, angles, center, detector_pad, recon_size, gpu_id
     )
 
     detY_size = data_shape[1]
@@ -140,6 +143,7 @@ def FBP3d_tomobar(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     filter_freq_cutoff: float = 0.35,
     recon_size: Optional[int] = None,
     recon_mask_radius: Optional[float] = 0.95,
@@ -159,6 +163,8 @@ def FBP3d_tomobar(
         An array of angles given in radians.
     center : float, optional
         The center of rotation (CoR).
+    detector_pad : int
+        Detector width padding with edge values to remove circle/arc type artifacts in the reconstruction.
     filter_freq_cutoff : float
         Cutoff frequency parameter for the SINC filter, the lower values produce better contrast but noisy reconstruction.
     recon_size : int, optional
@@ -182,7 +188,7 @@ def FBP3d_tomobar(
     data = data_checker(data, verbosity=True, method_name="FBP3d_tomobar")
 
     RecToolsCP = _instantiate_direct_recon_class(
-        data, angles, center, recon_size, gpu_id
+        data, angles, center, detector_pad, recon_size, gpu_id
     )
 
     reconstruction = RecToolsCP.FBP(
@@ -200,6 +206,7 @@ def LPRec3d_tomobar(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     filter_type: str = "shepp",
     filter_freq_cutoff: float = 1.0,
     recon_size: Optional[int] = None,
@@ -219,6 +226,8 @@ def LPRec3d_tomobar(
         An array of angles given in radians.
     center : float, optional
         The center of rotation (CoR).
+    detector_pad : int
+        Detector width padding with edge values to remove circle/arc type artifacts in the reconstruction.
     filter_type : str
         Filter type, the accepted strings are: none, ramp, shepp, cosine, cosine2, hamming, hann, parzen.
     filter_freq_cutoff : float
@@ -242,7 +251,9 @@ def LPRec3d_tomobar(
 
     data = data_checker(data, verbosity=True, method_name="LPRec3d_tomobar")
 
-    RecToolsCP = _instantiate_direct_recon_class(data, angles, center, recon_size, 0)
+    RecToolsCP = _instantiate_direct_recon_class(
+        data, angles, center, detector_pad, recon_size, 0
+    )
 
     reconstruction = RecToolsCP.FOURIER_INV(
         _take_neg_log(data) if neglog else data,
@@ -260,6 +271,7 @@ def SIRT3d_tomobar(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     recon_size: Optional[int] = None,
     iterations: Optional[int] = 300,
     nonnegativity: Optional[bool] = True,
@@ -280,6 +292,8 @@ def SIRT3d_tomobar(
         An array of angles given in radians.
     center : float, optional
         The center of rotation (CoR).
+    detector_pad : int
+        Detector width padding with edge values to remove circle/arc type artifacts in the reconstruction.
     recon_size : int, optional
         The [recon_size, recon_size] shape of the reconstructed slice in pixels.
         By default (None), the reconstructed size will be the dimension of the horizontal detector.
@@ -304,6 +318,7 @@ def SIRT3d_tomobar(
         data,
         angles,
         center,
+        detector_pad,
         recon_size,
         gpu_id,
         datafidelity="LS",
@@ -327,6 +342,7 @@ def CGLS3d_tomobar(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     recon_size: Optional[int] = None,
     iterations: Optional[int] = 20,
     nonnegativity: Optional[bool] = True,
@@ -347,6 +363,8 @@ def CGLS3d_tomobar(
         An array of angles given in radians.
     center : float, optional
         The center of rotation (CoR).
+    detector_pad : int
+        Detector width padding with edge values to remove circle/arc type artifacts in the reconstruction.
     recon_size : int, optional
         The [recon_size, recon_size] shape of the reconstructed slice in pixels.
         By default (None), the reconstructed size will be the dimension of the horizontal detector.
@@ -368,7 +386,7 @@ def CGLS3d_tomobar(
     data = data_checker(data, verbosity=True, method_name="CGLS3d_tomobar")
 
     RecToolsCP = _instantiate_iterative_recon_class(
-        data, angles, center, recon_size, gpu_id, datafidelity="LS"
+        data, angles, center, detector_pad, recon_size, gpu_id, datafidelity="LS"
     )
 
     _data_ = {
@@ -386,6 +404,7 @@ def _instantiate_direct_recon_class(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     recon_size: Optional[int] = None,
     gpu_id: int = 0,
 ) -> Type:
@@ -395,6 +414,7 @@ def _instantiate_direct_recon_class(
         data (cp.ndarray): data array
         angles (np.ndarray): angles
         center (Optional[float], optional): center of recon. Defaults to None.
+        detector_pad (int): Detector width padding. Defaults to 0.
         recon_size (Optional[int], optional): recon_size. Defaults to None.
         gpu_id (int, optional): gpu ID. Defaults to 0.
 
@@ -407,6 +427,7 @@ def _instantiate_direct_recon_class(
         recon_size = data.shape[2]
     RecToolsCP = RecToolsDIRCuPy(
         DetectorsDimH=data.shape[2],  # Horizontal detector dimension
+        DetectorsDimH_pad=detector_pad,  # padding for horizontal detector
         DetectorsDimV=data.shape[1],  # Vertical detector dimension (3D case)
         CenterRotOffset=data.shape[2] / 2
         - center
@@ -423,6 +444,7 @@ def _instantiate_direct_recon2d_class(
     data: np.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     recon_size: Optional[int] = None,
     gpu_id: int = 0,
 ) -> Type:
@@ -432,6 +454,7 @@ def _instantiate_direct_recon2d_class(
         data (cp.ndarray): data array
         angles (np.ndarray): angles
         center (Optional[float], optional): center of recon. Defaults to None.
+        detector_pad (int): Detector width padding. Defaults to 0.
         recon_size (Optional[int], optional): recon_size. Defaults to None.
         gpu_id (int, optional): gpu ID. Defaults to 0.
 
@@ -444,6 +467,7 @@ def _instantiate_direct_recon2d_class(
         recon_size = data.shape[2]
     RecTools = RecToolsDIR(
         DetectorsDimH=data.shape[2],  # Horizontal detector dimension
+        DetectorsDimH_pad=detector_pad,  # padding for horizontal detector
         DetectorsDimV=None,  # 2d case
         CenterRotOffset=data.shape[2] / 2
         - center
@@ -459,6 +483,7 @@ def _instantiate_iterative_recon_class(
     data: cp.ndarray,
     angles: np.ndarray,
     center: Optional[float] = None,
+    detector_pad: int = 0,
     recon_size: Optional[int] = None,
     gpu_id: int = 0,
     datafidelity: str = "LS",
@@ -469,6 +494,7 @@ def _instantiate_iterative_recon_class(
         data (cp.ndarray): data array
         angles (np.ndarray): angles
         center (Optional[float], optional): center of recon. Defaults to None.
+        detector_pad (int): Detector width padding. Defaults to 0.
         recon_size (Optional[int], optional): recon_size. Defaults to None.
         datafidelity (str, optional): Data fidelity
         gpu_id (int, optional): gpu ID. Defaults to 0.
@@ -482,6 +508,7 @@ def _instantiate_iterative_recon_class(
         recon_size = data.shape[2]
     RecToolsCP = RecToolsIRCuPy(
         DetectorsDimH=data.shape[2],  # Horizontal detector dimension
+        DetectorsDimH_pad=detector_pad,  # padding for horizontal detector
         DetectorsDimV=data.shape[1],  # Vertical detector dimension (3D case)
         CenterRotOffset=data.shape[2] / 2
         - center

tests/test_recon/test_algorithm.py

@@ -22,6 +22,7 @@ def test_reconstruct_FBP_2d_astra(data, flats, darks, ensure_clean_memory):
         cp.asnumpy(normalised_data),
         np.linspace(0.0 * np.pi / 180.0, 180.0 * np.pi / 180.0, data.shape[0]),
         79.5,
+        0,
         filter_type="shepp-logan",
         filter_parameter=None,
         filter_d=2.0,
@@ -35,6 +36,28 @@ def test_reconstruct_FBP_2d_astra(data, flats, darks, ensure_clean_memory):
     assert recon_data.shape == (recon_size, 128, recon_size)
 
 
+def test_reconstruct_FBP_2d_astra_pad(data, flats, darks, ensure_clean_memory):
+    normalised_data = normalize_cupy(data, flats, darks, cutoff=10, minus_log=True)
+    recon_size = 150
+
+    recon_data = FBP2d_astra(
+        cp.asnumpy(normalised_data),
+        np.linspace(0.0 * np.pi / 180.0, 180.0 * np.pi / 180.0, data.shape[0]),
+        79.5,
+        20,
+        filter_type="shepp-logan",
+        filter_parameter=None,
+        filter_d=2.0,
+        recon_size=recon_size,
+        recon_mask_radius=0.9,
+    )
+    assert recon_data.flags.c_contiguous
+    assert_allclose(np.mean(recon_data), 0.0020, atol=1e-04)
+    assert_allclose(np.mean(recon_data, axis=(0, 2)).sum(), 0.265565, rtol=1e-05)
+    assert recon_data.dtype == np.float32
+    assert recon_data.shape == (recon_size, 128, recon_size)
+
+
 def test_reconstruct_FBP3d_tomobar_1(data, flats, darks, ensure_clean_memory):
     recon_data = FBP3d_tomobar(
         normalize_cupy(data, flats, darks, cutoff=10, minus_log=True),
@@ -95,6 +118,7 @@ def test_reconstruct_FBP3d_tomobar_3(data, flats, darks, ensure_clean_memory):
         normalize_cupy(data, flats, darks, cutoff=20.5, minus_log=False),
         np.linspace(5.0 * np.pi / 360.0, 180.0 * np.pi / 360.0, data.shape[0]),
         79,  # center
+        0,  # detector pad
         1.1,  # filter_freq_cutoff
         210,  # recon_size
         0.9,  # recon_mask_radius
@@ -109,11 +133,48 @@ def test_reconstruct_FBP3d_tomobar_3(data, flats, darks, ensure_clean_memory):
     assert recon_data.shape == (210, 128, 210)
 
 
+def test_reconstruct_FBP3d_tomobar_3_pad(data, flats, darks, ensure_clean_memory):
+    recon_data = FBP3d_tomobar(
+        normalize_cupy(data, flats, darks, cutoff=20.5, minus_log=False),
+        np.linspace(5.0 * np.pi / 360.0, 180.0 * np.pi / 360.0, data.shape[0]),
+        79,  # center
+        20,  # detector pad
+        1.1,  # filter_freq_cutoff
+        210,  # recon_size
+        0.9,  # recon_mask_radius
+    )
+
+    recon_data = recon_data.get()
+    assert_allclose(np.mean(recon_data), -0.000392, atol=1e-6)
+    assert_allclose(
+        np.mean(recon_data, axis=(0, 2)).sum(), -0.050226, rtol=1e-06, atol=1e-5
+    )
+    assert recon_data.dtype == np.float32
+    assert recon_data.shape == (210, 128, 210)
+
+
 def test_reconstruct_LPRec3d_tomobar_1(data, flats, darks, ensure_clean_memory):
     recon_data = LPRec3d_tomobar(
         data=normalize_cupy(data, flats, darks, cutoff=10, minus_log=True),
         angles=np.linspace(0.0 * np.pi / 180.0, 180.0 * np.pi / 180.0, data.shape[0]),
         center=79.5,
+        detector_pad=0,
+        recon_size=130,
+        recon_mask_radius=0.95,
+    )
+    assert recon_data.flags.c_contiguous
+    recon_data = recon_data.get()
+    assert_allclose(np.mean(recon_data), 0.007, atol=1e-3)
+    assert recon_data.dtype == np.float32
+    assert recon_data.shape == (130, 128, 130)
+
+
+def test_reconstruct_LPRec3d_tomobar_1_pad(data, flats, darks, ensure_clean_memory):
+    recon_data = LPRec3d_tomobar(
+        data=normalize_cupy(data, flats, darks, cutoff=10, minus_log=True),
+        angles=np.linspace(0.0 * np.pi / 180.0, 180.0 * np.pi / 180.0, data.shape[0]),
+        center=79.5,
+        detector_pad=10,
         recon_size=130,
         recon_mask_radius=0.95,
     )
@@ -165,10 +226,11 @@ def test_FBP3d_tomobar_performance(ensure_clean_memory):
     data = cp.asarray(data_host, dtype=np.float32)
     angles = np.linspace(0.0 * np.pi / 180.0, 180.0 * np.pi / 180.0, data.shape[0])
     cor = 79.5
+    det_pad = 0
     filter_freq_cutoff = 1.1
 
     # cold run first
-    FBP3d_tomobar(data, angles, cor, filter_freq_cutoff)
+    FBP3d_tomobar(data, angles, cor, det_pad, filter_freq_cutoff)
     dev.synchronize()
 
     start = time.perf_counter_ns()