"""The CT module automatically analyzes DICOM images of a CatPhan 504, 503, 600, Quart DVT, or ACR phantoms acquired when doing CBCT or CT quality assurance.
It can load a folder or zip file that the images are in and automatically correct for translational and rotational errors.
It can analyze the HU regions and image scaling (CTP404), the high-contrast line pairs (CTP528) to calculate the modulation transfer function (MTF),
the HU uniformity (CTP486), and Low Contrast (CTP515) on the corresponding slices.
For ACR and Quart phantoms, the equivalent sections are analyzed where applicable even though each module does not have an explicit name.
Where intuitive similarities between the phantoms exist, the library usage is the same.
Features:
* **Automatic phantom registration** - Your phantom can be tilted, rotated, or translated--pylinac will automatically register the phantom.
* **Automatic testing of all major modules** - Major modules are automatically registered and analyzed.
* **Any scan protocol** - Scan your CatPhan with any protocol; even scan it in a regular CT scanner.
Any field size or field extent is allowed.
"""
from __future__ import annotations
import io
import itertools
import os
import textwrap
import webbrowser
import zipfile
from functools import cached_property
from io import BytesIO
from os import path as osp
from pathlib import Path
from typing import BinaryIO, Callable, Sequence
import matplotlib.pyplot as plt
import numpy as np
from matplotlib.axes import Axes
from py_linq import Enumerable
from pydantic import BaseModel
from scipy import ndimage
from skimage import draw, filters, measure, segmentation
from skimage.measure._regionprops import RegionProperties
from .core import image, pdf
from .core.contrast import Contrast
from .core.geometry import Line, Point
from .core.image import ArrayImage, DicomImageStack, ImageLike, z_position
from .core.io import TemporaryZipDirectory, get_url, retrieve_demo_file
from .core.mtf import MTF
from .core.nps import (
average_power,
max_frequency,
noise_power_spectrum_1d,
noise_power_spectrum_2d,
)
from .core.profile import CollapsedCircleProfile, FWXMProfile
from .core.roi import DiskROI, LowContrastDiskROI, RectangleROI
from .core.utilities import ResultBase, ResultsDataMixin
from .settings import get_dicom_cmap
# The ramp angle ratio is from the Catphan manual ("Scan slice geometry" section)
# and represents the fact that the wire is at an oblique angle (23°), making it appear
# longer than it is if it were normal or perpendicular to the z (imaging) axis. This ratio
# fixes the length to represent it as if it were perpendicular to the imaging axis.
RAMP_ANGLE_RATIO = 0.42
AIR = -1000
PMP = -196
LDPE = -104
POLY = -47
ACRYLIC = 115
DELRIN = 365
TEFLON = 1000
BONE_20 = 237
BONE_50 = 725
WATER = 0
[docs]
class ROIResult(BaseModel):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
name: str #:
value: float #:
stdev: float #:
difference: float | None #:
nominal_value: float | None #:
passed: bool | None #:
[docs]
class CTP404Result(BaseModel):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
offset: int #:
low_contrast_visibility: float #:
thickness_passed: bool #:
measured_slice_thickness_mm: float #:
thickness_num_slices_combined: int #:
geometry_passed: bool #:
avg_line_distance_mm: float #:
line_distances_mm: list[float] #:
hu_linearity_passed: bool #:
hu_tolerance: float #:
hu_rois: dict #:
[docs]
class CTP486Result(BaseModel):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
uniformity_index: float #:
integral_non_uniformity: float #:
nps_avg_power: float
nps_max_freq: float
passed: bool #:
rois: dict #:
[docs]
class CTP515Result(BaseModel):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
cnr_threshold: float #:
num_rois_seen: int #:
roi_settings: dict #:
roi_results: dict #:
[docs]
class CTP528Result(BaseModel):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
start_angle_radians: float #:
mtf_lp_mm: dict #:
roi_settings: dict #:
[docs]
class CatphanResult(ResultBase):
"""This class should not be called directly. It is returned by the ``results_data()`` method.
It is a dataclass under the hood and thus comes with all the dunder magic.
Use the following attributes as normal class attributes."""
catphan_model: str #:
catphan_roll_deg: float #:
origin_slice: int #:
num_images: int #:
ctp404: CTP404Result #:
ctp486: CTP486Result | None = None #:
ctp528: CTP528Result | None = None #:
ctp515: CTP515Result | None = None #:
[docs]
class HUDiskROI(DiskROI):
"""An HU ROI object. Represents a circular area measuring either HU sample (Air, Poly, ...)
or HU uniformity (bottom, left, ...).
"""
def __init__(
self,
array: np.ndarray | ArrayImage,
angle: float,
roi_radius: float,
dist_from_center: float,
phantom_center: tuple | Point,
nominal_value: float | None = None,
tolerance: float | None = None,
background_mean: float | None = None,
background_std: float | None = None,
):
"""
Parameters
----------
nominal_value
The nominal pixel value of the HU ROI.
tolerance
The roi pixel value tolerance.
"""
super().__init__(array, angle, roi_radius, dist_from_center, phantom_center)
self.nominal_val = nominal_value
self.tolerance = tolerance
@property
def value_diff(self) -> float:
"""The difference in HU between measured and nominal."""
return self.pixel_value - self.nominal_val
@property
def passed(self) -> bool:
"""Boolean specifying if ROI pixel value was within tolerance of the nominal value."""
if self.tolerance:
return abs(self.value_diff) <= self.tolerance
else:
return True
@property
def plot_color(self) -> str:
"""Return one of two colors depending on if ROI passed."""
return "green" if self.passed else "red"
[docs]
class ThicknessROI(RectangleROI):
"""A rectangular ROI that measures the angled wire rod in the HU linearity slice which determines slice thickness."""
@cached_property
def long_profile(self) -> FWXMProfile:
"""The profile along the axis perpendicular to ramped wire."""
img = image.load(self.pixel_array)
img.filter(size=1, kind="gaussian")
return FWXMProfile(values=img.array.max(axis=np.argmin(img.shape)))
@cached_property
def wire_fwhm(self) -> float:
"""The FWHM of the wire in pixels."""
return self.long_profile.field_width_px
@property
def plot_color(self) -> str:
"""The plot color."""
return "blue"
[docs]
class Slice:
"""Base class for analyzing specific slices of a CBCT dicom set."""
def __init__(
self,
catphan,
slice_num: int | None = None,
combine: bool = True,
combine_method: str = "mean",
num_slices: int = 0,
clear_borders: bool = True,
original_image: ImageLike | None = None,
):
"""
Parameters
----------
catphan : :class:`~pylinac.cbct.CatPhanBase` instance.
The catphan instance.
slice_num : int
The slice number of the DICOM array desired. If None, will use the ``slice_num`` property of subclass.
combine : bool
If True, combines the slices +/- ``num_slices`` around the slice of interest to improve signal/noise.
combine_method : {'mean', 'max'}
How to combine the slices if ``combine`` is True.
num_slices : int
The number of slices on either side of the nominal slice to combine to improve signal/noise; only
applicable if ``combine`` is True.
clear_borders : bool
If True, clears the borders of the image to remove any ROIs that may be present.
original_image : :class:`~pylinac.core.image.Image` or None
The array of the slice. This is a bolt-on parameter for optimization.
Leaving as None is fine, but can increase analysis speed if 1) this image is passed and
2) there is no combination of slices happening, which is most of the time.
"""
if slice_num is not None:
self.slice_num = slice_num
if combine and num_slices > 0:
array = combine_surrounding_slices(
catphan.dicom_stack,
self.slice_num,
mode=combine_method,
slices_plusminus=num_slices,
)
elif original_image is not None:
array = original_image
else:
array = catphan.dicom_stack[self.slice_num].array
self.image = image.load(array)
self.catphan_size = catphan.catphan_size
self.mm_per_pixel = catphan.mm_per_pixel
self.clear_borders = clear_borders
if catphan._phantom_center_func:
self._phantom_center_func = catphan._phantom_center_func
@property
def __getitem__(self, item):
return self.image.array[item]
@cached_property
def phantom_roi(self) -> RegionProperties:
"""Get the Scikit-Image ROI of the phantom
The image is analyzed to see if:
1) the CatPhan is even in the image (if there were any ROIs detected)
2) an ROI is within the size criteria of the catphan
3) the ROI area that is filled compared to the bounding box area is close to that of a circle
"""
# convert the slice to binary and label ROIs
edges = filters.scharr(self.image.as_type(float))
if np.max(edges) < 0.1:
raise ValueError(
"No edges were found in the image that look like the phantom"
)
larr, regionprops, num_roi = get_regions(
self, fill_holes=True, threshold="otsu", clear_borders=self.clear_borders
)
# check that there is at least 1 ROI
if num_roi < 1 or num_roi is None:
raise ValueError(
f"The number of ROIs detected {num_roi} was not the number expected (1)"
)
catphan_region = sorted(
regionprops, key=lambda x: np.abs(x.filled_area - self.catphan_size)
)[0]
if (self.catphan_size * 1.3 < catphan_region.filled_area) or (
catphan_region.filled_area < self.catphan_size / 1.3
):
raise ValueError("Unable to find ROI of expected size of the phantom")
return catphan_region
[docs]
def is_phantom_in_view(self) -> bool:
"""Whether the phantom appears to be within the slice."""
try:
self.phantom_roi
return True
except ValueError:
return False
@property
def phan_center(self) -> Point:
"""Determine the location of the center of the phantom."""
x = self._phantom_center_func[0](self.slice_num)
y = self._phantom_center_func[1](self.slice_num)
return Point(x=x, y=y)
[docs]
class CatPhanModule(Slice):
"""Base class for a CTP module."""
common_name: str = ""
combine_method: str = "mean"
num_slices: int = 0
roi_settings: dict = {}
background_roi_settings: dict = {}
roi_dist_mm = float
roi_radius_mm = float
rois: dict = {} # dicts of HUDiskROIs
background_rois: dict = {} # dict of HUDiskROIs; possibly empty
window_min: int | None = None # plt visualization
window_max: int | None = None # plt visualization
def __init__(
self,
catphan,
tolerance: float | None = None,
offset: int = 0,
clear_borders: bool = True,
):
self.model = ""
self._offset = offset
self.origin_slice = catphan.origin_slice
self.tolerance = tolerance
self.slice_thickness = catphan.dicom_stack.metadata.SliceThickness
self.slice_spacing = catphan.dicom_stack.slice_spacing
self.catphan_roll = catphan.catphan_roll
self.mm_per_pixel = catphan.mm_per_pixel
self.rois: dict[str, HUDiskROI] = {}
self.background_rois: dict[str, HUDiskROI] = {}
Slice.__init__(
self,
catphan,
combine_method=self.combine_method,
num_slices=self.num_slices,
clear_borders=clear_borders,
)
self._convert_units_in_settings()
self.preprocess(catphan)
self._setup_rois()
def _convert_units_in_settings(self) -> None:
setting_groups = [
getattr(self, attr) for attr in dir(self) if attr.endswith("roi_settings")
]
for roi_settings in setting_groups:
for roi, settings in roi_settings.items():
if isinstance(settings, dict):
if settings.get("distance") is not None:
settings["distance_pixels"] = (
settings["distance"] / self.mm_per_pixel
)
if settings.get("angle") is not None:
settings["angle_corrected"] = (
settings["angle"] + self.catphan_roll
)
if settings.get("radius") is not None:
settings["radius_pixels"] = (
settings["radius"] / self.mm_per_pixel
)
if settings.get("width") is not None:
settings["width_pixels"] = settings["width"] / self.mm_per_pixel
if settings.get("height") is not None:
settings["height_pixels"] = (
settings["height"] / self.mm_per_pixel
)
[docs]
def preprocess(self, catphan):
"""A preprocessing step before analyzing the CTP module.
Parameters
----------
catphan : `~pylinac.cbct.CatPhanBase` instance.
"""
pass
@property
def slice_num(self) -> int:
"""The slice number of the spatial resolution module.
Returns
-------
float
"""
return int(self.origin_slice + round(self._offset / self.slice_spacing))
def _setup_rois(self) -> None:
for name, setting in self.background_roi_settings.items():
self.background_rois[name] = HUDiskROI(
self.image,
setting["angle_corrected"],
setting["radius_pixels"],
setting["distance_pixels"],
self.phan_center,
)
if self.background_rois:
background_mean = np.mean(
[roi.pixel_value for roi in self.background_rois.values()]
)
background_std = np.std(
[roi.pixel_value for roi in self.background_rois.values()]
)
else:
background_mean = None
background_std = None
for name, setting in self.roi_settings.items():
nominal_value = setting.get("value", 0)
self.rois[name] = HUDiskROI(
self.image,
setting["angle_corrected"],
setting["radius_pixels"],
setting["distance_pixels"],
self.phan_center,
nominal_value,
self.tolerance,
background_mean=background_mean,
background_std=background_std,
)
[docs]
def plot_rois(self, axis: plt.Axes) -> None:
"""Plot the ROIs to the axis."""
for roi in self.rois.values():
roi.plot2axes(axis, edgecolor=roi.plot_color)
for roi in self.background_rois.values():
roi.plot2axes(axis, edgecolor="blue")
[docs]
def plot(self, axis: plt.Axes):
"""Plot the image along with ROIs to an axis"""
axis.imshow(
self.image.array,
cmap=get_dicom_cmap(),
vmin=self.window_min,
vmax=self.window_max,
)
self.plot_rois(axis)
axis.autoscale(tight=True)
axis.set_title(self.common_name)
axis.axis("off")
@property
def roi_vals_as_str(self) -> str:
return ", ".join(
f"{name}: {roi.pixel_value}" for name, roi in self.rois.items()
)
[docs]
class CTP404CP504(CatPhanModule):
"""Class for analysis of the HU linearity, geometry, and slice thickness regions of the CTP404."""
attr_name = "ctp404"
common_name = "HU Linearity"
roi_dist_mm = 58.7
roi_radius_mm = 5
roi_settings = {
"Air": {
"value": AIR,
"angle": -90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"PMP": {
"value": PMP,
"angle": -120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"LDPE": {
"value": LDPE,
"angle": 180,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Poly": {
"value": POLY,
"angle": 120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Acrylic": {
"value": ACRYLIC,
"angle": 60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Delrin": {
"value": DELRIN,
"angle": 0,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Teflon": {
"value": TEFLON,
"angle": -60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
}
background_roi_settings = {
"1": {"angle": -30, "distance": roi_dist_mm, "radius": roi_radius_mm},
"2": {"angle": -150, "distance": roi_dist_mm, "radius": roi_radius_mm},
"3": {"angle": -210, "distance": roi_dist_mm, "radius": roi_radius_mm},
"4": {"angle": 30, "distance": roi_dist_mm, "radius": roi_radius_mm},
}
# thickness
thickness_roi_height = 40
thickness_roi_width = 10
thickness_roi_distance_mm = 38
thickness_roi_settings = {
"Left": {
"angle": 180,
"width": thickness_roi_width,
"height": thickness_roi_height,
"distance": thickness_roi_distance_mm,
},
"Bottom": {
"angle": 90,
"width": thickness_roi_height,
"height": thickness_roi_width,
"distance": thickness_roi_distance_mm,
},
"Right": {
"angle": 0,
"width": thickness_roi_width,
"height": thickness_roi_height,
"distance": thickness_roi_distance_mm,
},
"Top": {
"angle": -90,
"width": thickness_roi_height,
"height": thickness_roi_width,
"distance": thickness_roi_distance_mm,
},
}
# geometry
geometry_roi_size_mm = 35
geometry_roi_settings = {
"Top-Horizontal": (0, 1),
"Bottom-Horizontal": (2, 3),
"Left-Vertical": (0, 2),
"Right-Vertical": (1, 3),
}
pad: str | int
thickness_image: Slice
def __init__(
self,
catphan,
offset: int,
hu_tolerance: float,
thickness_tolerance: float,
scaling_tolerance: float,
clear_borders: bool = True,
thickness_slice_straddle: str | int = "auto",
expected_hu_values: dict[str, float | int] | None = None,
):
"""
Parameters
----------
catphan : `~pylinac.cbct.CatPhanBase` instance.
offset : int
hu_tolerance : float
thickness_tolerance : float
scaling_tolerance : float
clear_borders : bool
"""
self.mm_per_pixel = catphan.mm_per_pixel
self.hu_tolerance = hu_tolerance
self.thickness_tolerance = thickness_tolerance
self.scaling_tolerance = scaling_tolerance
self.thickness_rois = {}
self.lines = {}
self.thickness_slice_straddle = thickness_slice_straddle
self.expected_hu_values = expected_hu_values
super().__init__(
catphan, tolerance=hu_tolerance, offset=offset, clear_borders=clear_borders
)
[docs]
def preprocess(self, catphan) -> None:
# for the thickness analysis image, combine thin slices or just use one slice if slices are thick
if (
isinstance(self.thickness_slice_straddle, str)
and self.thickness_slice_straddle.lower() == "auto"
):
if float(catphan.dicom_stack.metadata.SliceThickness) < 3.5:
self.pad = 1
else:
self.pad = 0
else:
self.pad = self.thickness_slice_straddle
self.thickness_image = Slice(
catphan,
combine_method="mean",
num_slices=self.num_slices + self.pad,
slice_num=self.slice_num,
clear_borders=self.clear_borders,
).image
def _replace_hu_values(self):
"""Possibly replace the HU values in the ROI settings with the expected values if the key is present."""
if self.expected_hu_values is not None:
for name, value in self.expected_hu_values.items():
if name in self.roi_settings:
self.roi_settings[name]["value"] = value
def _setup_rois(self) -> None:
self._replace_hu_values()
super()._setup_rois()
self._setup_thickness_rois()
self._setup_geometry_rois()
def _setup_thickness_rois(self) -> None:
for name, setting in self.thickness_roi_settings.items():
self.thickness_rois[name] = ThicknessROI(
self.thickness_image,
setting["width_pixels"],
setting["height_pixels"],
setting["angle_corrected"],
setting["distance_pixels"],
self.phan_center,
)
def _setup_geometry_rois(self) -> None:
boxsize = self.geometry_roi_size_mm / self.mm_per_pixel
xbounds = (int(self.phan_center.x - boxsize), int(self.phan_center.x + boxsize))
ybounds = (int(self.phan_center.y - boxsize), int(self.phan_center.y + boxsize))
geo_img = self.image[ybounds[0] : ybounds[1], xbounds[0] : xbounds[1]]
larr, regionprops, num_roi = get_regions(
geo_img, fill_holes=True, clear_borders=False
)
# check that there is at least 1 ROI
if num_roi < 4:
raise ValueError("Unable to locate the Geometric nodes")
elif num_roi > 4:
regionprops = sorted(
regionprops, key=lambda x: x.filled_area, reverse=True
)[:4]
sorted_regions = sorted(
regionprops, key=lambda x: (2 * x.centroid[0] + x.centroid[1])
)
centers = [
Point(
r.weighted_centroid[1] + xbounds[0], r.weighted_centroid[0] + ybounds[0]
)
for r in sorted_regions
]
# setup the geometric lines
for name, order in self.geometry_roi_settings.items():
self.lines[name] = GeometricLine(
centers[order[0]],
centers[order[1]],
self.mm_per_pixel,
self.scaling_tolerance,
)
@property
def lcv(self) -> float:
"""The low-contrast visibility"""
return (
2
* abs(self.rois["LDPE"].pixel_value - self.rois["Poly"].pixel_value)
/ (self.rois["LDPE"].std + self.rois["Poly"].std)
)
[docs]
def plot_linearity(
self, axis: plt.Axes | None = None, plot_delta: bool = True
) -> tuple:
"""Plot the HU linearity values to an axis.
Parameters
----------
axis : None, matplotlib.Axes
The axis to plot the values on. If None, will create a new figure.
plot_delta : bool
Whether to plot the actual measured HU values (False), or the difference from nominal (True).
"""
nominal_x_values = [roi.nominal_val for roi in self.rois.values()]
if axis is None:
fig, axis = plt.subplots()
if plot_delta:
values = [roi.value_diff for roi in self.rois.values()]
nominal_measurements = [0] * len(values)
ylabel = "HU Delta"
else:
values = [roi.pixel_value for roi in self.rois.values()]
nominal_measurements = nominal_x_values
ylabel = "Measured Values"
points = axis.plot(nominal_x_values, values, "g+", markersize=15, mew=2)
axis.plot(nominal_x_values, nominal_measurements)
axis.plot(
nominal_x_values, np.array(nominal_measurements) + self.hu_tolerance, "r--"
)
axis.plot(
nominal_x_values, np.array(nominal_measurements) - self.hu_tolerance, "r--"
)
axis.margins(0.05)
axis.grid(True)
axis.set_xlabel("Nominal Values")
axis.set_ylabel(ylabel)
axis.set_title("HU linearity")
return points
@property
def passed_hu(self) -> bool:
"""Boolean specifying whether all the ROIs passed within tolerance."""
return all(roi.passed for roi in self.rois.values())
[docs]
def plot_rois(self, axis: plt.Axes) -> None:
"""Plot the ROIs onto the image, as well as the background ROIs"""
# plot HU linearity ROIs
super().plot_rois(axis)
# plot thickness ROIs
for roi in self.thickness_rois.values():
roi.plot2axes(axis, edgecolor="blue")
# plot geometry lines
for line in self.lines.values():
line.plot2axes(axis, color=line.pass_fail_color)
@property
def passed_thickness(self) -> bool:
"""Whether the slice thickness was within tolerance from nominal."""
return (
self.slice_thickness - self.thickness_tolerance
< self.meas_slice_thickness
< self.slice_thickness + self.thickness_tolerance
)
@property
def meas_slice_thickness(self) -> float:
"""The average slice thickness for the 4 wire measurements in mm."""
return np.mean(
sorted(
roi.wire_fwhm * self.mm_per_pixel * RAMP_ANGLE_RATIO
for roi in self.thickness_rois.values()
)
) / (1 + 2 * self.pad)
@property
def avg_line_length(self) -> float:
return float(np.mean([line.length_mm for line in self.lines.values()]))
@property
def passed_geometry(self) -> bool:
"""Returns whether all the line lengths were within tolerance."""
return all(line.passed for line in self.lines.values())
[docs]
class CTP404CP503(CTP404CP504):
"""Alias for namespace consistency"""
pass
[docs]
class CTP404CP600(CTP404CP504):
roi_dist_mm = 58.7
roi_radius_mm = 5
roi_settings = {
"Air": {
"value": AIR,
"angle": 90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"PMP": {
"value": PMP,
"angle": 60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"LDPE": {
"value": LDPE,
"angle": 0,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Poly": {
"value": POLY,
"angle": -60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Acrylic": {
"value": ACRYLIC,
"angle": -120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Delrin": {
"value": DELRIN,
"angle": -180,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Teflon": {
"value": TEFLON,
"angle": 120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Vial": {
"value": WATER,
"angle": -90,
"distance": roi_dist_mm,
"radius": roi_radius_mm
- 1, # the vial sits inside the ROI and needs some clearance
},
}
def _setup_rois(self) -> None:
"""For the 600, the top ROI is an optional water vial slot. If the HU is near water we leave it, otherwise we remove it so as not to flag false failures"""
super()._setup_rois()
if self.rois["Vial"].pixel_value < -500: # closer to air than water
self.rois.pop("Vial")
[docs]
class CTP404CP604(CTP404CP504):
roi_dist_mm = 58.7
roi_radius_mm = 5
roi_settings = {
"Air": {
"value": AIR,
"angle": -90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"PMP": {
"value": PMP,
"angle": -120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"50% Bone": {
"value": BONE_50,
"angle": -150,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"LDPE": {
"value": LDPE,
"angle": 180,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Poly": {
"value": POLY,
"angle": 120,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Acrylic": {
"value": ACRYLIC,
"angle": 60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"20% Bone": {
"value": BONE_20,
"angle": 30,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Delrin": {
"value": DELRIN,
"angle": 0,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Teflon": {
"value": TEFLON,
"angle": -60,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
}
background_roi_settings = {
"1": {"angle": -30, "distance": roi_dist_mm, "radius": roi_radius_mm},
"2": {"angle": -210, "distance": roi_dist_mm, "radius": roi_radius_mm},
}
[docs]
class CTP486(CatPhanModule):
"""Class for analysis of the Uniformity slice of the CTP module. Measures 5 ROIs around the slice that
should all be close to the same value.
"""
attr_name = "ctp486"
common_name = "HU Uniformity"
roi_dist_mm = 53
roi_radius_mm = 10
nominal_value = 0
nps_rois: dict[str, RectangleROI]
roi_settings = {
"Top": {
"value": nominal_value,
"angle": -90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Right": {
"value": nominal_value,
"angle": 0,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Bottom": {
"value": nominal_value,
"angle": 90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Left": {
"value": nominal_value,
"angle": 180,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
},
"Center": {
"value": nominal_value,
"angle": 0,
"distance": 0,
"radius": roi_radius_mm,
},
}
[docs]
def plot_profiles(self, axis: plt.Axes | None = None) -> None:
"""Plot the horizontal and vertical profiles of the Uniformity slice.
Parameters
----------
axis : None, matplotlib.Axes
The axis to plot on; if None, will create a new figure.
"""
if axis is None:
fig, axis = plt.subplots()
horiz_data = self.image[int(self.phan_center.y), :]
vert_data = self.image[:, int(self.phan_center.x)]
axis.plot(horiz_data, "g", label="Horizontal")
axis.plot(vert_data, "b", label="Vertical")
axis.autoscale(tight=True)
axis.axhline(self.nominal_value + self.tolerance, color="r", linewidth=3)
axis.axhline(self.nominal_value - self.tolerance, color="r", linewidth=3)
axis.grid(True)
axis.set_ylabel("HU")
axis.legend(loc=8, fontsize="small", title="")
axis.set_title("Uniformity Profiles")
def _setup_rois(self) -> None:
"""Generate our NPS ROIs. They are just square versions of the existing ROIs."""
super()._setup_rois()
self.nps_rois = {}
for name, setting in self.roi_settings.items():
self.nps_rois[name] = RectangleROI(
array=self.image,
width=setting["radius_pixels"] * 2,
height=setting["radius_pixels"] * 2,
angle=setting["angle_corrected"],
dist_from_center=setting["distance_pixels"],
phantom_center=self.phan_center,
)
[docs]
def plot(self, axis: plt.Axes):
"""Plot the ROIs but also the noise power spectrum ROIs"""
for nps_roi in self.nps_rois.values():
nps_roi.plot2axes(axis, edgecolor="green", linestyle="-.")
super().plot(axis)
@property
def overall_passed(self) -> bool:
"""Boolean specifying whether all the ROIs passed within tolerance."""
return all(roi.passed for roi in self.rois.values())
@property
def uniformity_index(self) -> float:
"""The Uniformity Index. Elstrom et al equation 2. https://www.tandfonline.com/doi/pdf/10.3109/0284186X.2011.590525"""
center = self.rois["Center"]
uis = [
100 * ((roi.pixel_value - center.pixel_value) / (center.pixel_value + 1000))
for roi in self.rois.values()
]
abs_uis = np.abs(uis)
return uis[np.argmax(abs_uis)]
@property
def integral_non_uniformity(self) -> float:
"""The Integral Non-Uniformity. Elstrom et al equation 1. https://www.tandfonline.com/doi/pdf/10.3109/0284186X.2011.590525"""
maxhu = max(roi.pixel_value for roi in self.rois.values())
minhu = min(roi.pixel_value for roi in self.rois.values())
return (maxhu - minhu) / (maxhu + minhu + 2000)
@cached_property
def power_spectrum_2d(self) -> np.ndarray:
"""The power spectrum of the uniformity ROI."""
return noise_power_spectrum_2d(
pixel_size=self.mm_per_pixel,
rois=[r.pixel_array for r in self.nps_rois.values()],
)
@cached_property
def power_spectrum_1d(self) -> np.ndarray:
"""The 1D power spectrum of the uniformity ROI."""
return noise_power_spectrum_1d(self.power_spectrum_2d)
@property
def avg_noise_power(self) -> float:
"""The average noise power of the uniformity ROI."""
return average_power(self.power_spectrum_1d)
@property
def max_noise_power_frequency(self) -> float:
"""The frequency of the maximum noise power. 0 means no pattern."""
return max_frequency(self.power_spectrum_1d)
[docs]
class CTP528CP504(CatPhanModule):
"""Class for analysis of the Spatial Resolution slice of the CBCT dicom data set.
A collapsed circle profile is taken of the line-pair region. This profile is search for
peaks and valleys. The MTF is calculated from those peaks & valleys.
Attributes
----------
radius2linepairs_mm : float
The radius in mm to the line pairs.
"""
attr_name: str = "ctp528"
common_name: str = "Spatial Resolution"
radius2linepairs_mm = 47
combine_method: str = "max"
num_slices: int = 3
boundaries: tuple[float, ...] = (
0,
0.107,
0.173,
0.236,
0.286,
0.335,
0.387,
0.434,
0.479,
)
start_angle: float = np.pi
ccw: bool = True
roi_settings = {
"region 1": {
"start": boundaries[0],
"end": boundaries[1],
"num peaks": 2,
"num valleys": 1,
"peak spacing": 0.021,
"gap size (cm)": 0.5,
"lp/mm": 0.1,
},
"region 2": {
"start": boundaries[1],
"end": boundaries[2],
"num peaks": 3,
"num valleys": 2,
"peak spacing": 0.01,
"gap size (cm)": 0.25,
"lp/mm": 0.2,
},
"region 3": {
"start": boundaries[2],
"end": boundaries[3],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.006,
"gap size (cm)": 0.167,
"lp/mm": 0.3,
},
"region 4": {
"start": boundaries[3],
"end": boundaries[4],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.00557,
"gap size (cm)": 0.125,
"lp/mm": 0.4,
},
"region 5": {
"start": boundaries[4],
"end": boundaries[5],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.004777,
"gap size (cm)": 0.1,
"lp/mm": 0.5,
},
"region 6": {
"start": boundaries[5],
"end": boundaries[6],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.00398,
"gap size (cm)": 0.083,
"lp/mm": 0.6,
},
"region 7": {
"start": boundaries[6],
"end": boundaries[7],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.00358,
"gap size (cm)": 0.071,
"lp/mm": 0.7,
},
"region 8": {
"start": boundaries[7],
"end": boundaries[8],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.0027866,
"gap size (cm)": 0.063,
"lp/mm": 0.8,
},
}
def _setup_rois(self):
pass
def _convert_units_in_settings(self):
pass
@cached_property
def mtf(self) -> MTF:
"""The Relative MTF of the line pairs, normalized to the first region.
Returns
-------
dict
"""
maxs = list()
mins = list()
for key, value in self.roi_settings.items():
max_indices, max_values = self.circle_profile.find_peaks(
min_distance=value["peak spacing"],
max_number=value["num peaks"],
search_region=(value["start"], value["end"]),
)
# check that the right number of peaks were found before continuing, otherwise stop searching for regions
if len(max_values) != value["num peaks"]:
break
maxs.append(max_values.mean())
_, min_values = self.circle_profile.find_valleys(
min_distance=value["peak spacing"],
max_number=value["num valleys"],
search_region=(min(max_indices), max(max_indices)),
)
mins.append(min_values.mean())
if not maxs:
raise ValueError(
"Did not find any spatial resolution pairs to analyze. File an issue on github (https://github.com/jrkerns/pylinac/issues) if this is a valid dataset."
)
spacings = [roi["lp/mm"] for roi in self.roi_settings.values()]
mtf = MTF(lp_spacings=spacings, lp_maximums=maxs, lp_minimums=mins)
return mtf
@property
def radius2linepairs(self) -> float:
"""Radius from the phantom center to the line-pair region, corrected for pixel spacing."""
return self.radius2linepairs_mm / self.mm_per_pixel
[docs]
def plot_rois(self, axis: plt.Axes) -> None:
"""Plot the circles where the profile was taken within."""
self.circle_profile.plot2axes(axis, edgecolor="blue", plot_peaks=False)
@cached_property
def circle_profile(self) -> CollapsedCircleProfile:
"""Calculate the median profile of the Line Pair region.
Returns
-------
:class:`pylinac.core.profile.CollapsedCircleProfile` : A 1D profile of the Line Pair region.
"""
circle_profile = CollapsedCircleProfile(
self.phan_center,
self.radius2linepairs,
image_array=self.image,
start_angle=self.start_angle + np.deg2rad(self.catphan_roll),
width_ratio=0.04,
sampling_ratio=2,
ccw=self.ccw,
)
circle_profile.filter(0.001, kind="gaussian")
circle_profile.ground()
return circle_profile
[docs]
class CTP528CP604(CTP528CP504):
"""Alias for namespace consistency."""
pass
[docs]
class CTP528CP600(CTP528CP504):
start_angle = np.pi - 0.1
ccw = False
boundaries = (0, 0.127, 0.195, 0.255, 0.304, 0.354, 0.405, 0.453, 0.496)
roi_settings = {
"region 1": {
"start": boundaries[0],
"end": boundaries[1],
"num peaks": 2,
"num valleys": 1,
"peak spacing": 0.021,
"gap size (cm)": 0.5,
"lp/mm": 0.1,
},
"region 2": {
"start": boundaries[1],
"end": boundaries[2],
"num peaks": 3,
"num valleys": 2,
"peak spacing": 0.01,
"gap size (cm)": 0.25,
"lp/mm": 0.2,
},
"region 3": {
"start": boundaries[2],
"end": boundaries[3],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.006,
"gap size (cm)": 0.167,
"lp/mm": 0.3,
},
"region 4": {
"start": boundaries[3],
"end": boundaries[4],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.00557,
"gap size (cm)": 0.125,
"lp/mm": 0.4,
},
"region 5": {
"start": boundaries[4],
"end": boundaries[5],
"num peaks": 4,
"num valleys": 3,
"peak spacing": 0.004777,
"gap size (cm)": 0.1,
"lp/mm": 0.5,
},
"region 6": {
"start": boundaries[5],
"end": boundaries[6],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.00398,
"gap size (cm)": 0.083,
"lp/mm": 0.6,
},
"region 7": {
"start": boundaries[6],
"end": boundaries[7],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.00358,
"gap size (cm)": 0.071,
"lp/mm": 0.7,
},
"region 8": {
"start": boundaries[7],
"end": boundaries[8],
"num peaks": 5,
"num valleys": 4,
"peak spacing": 0.0027866,
"gap size (cm)": 0.063,
"lp/mm": 0.8,
},
}
[docs]
class CTP528CP503(CTP528CP504):
start_angle = 0
ccw = False
boundaries = (0, 0.111, 0.176, 0.240, 0.289, 0.339, 0.390, 0.436, 0.481)
[docs]
class GeometricLine(Line):
"""Represents a line connecting two nodes/ROIs on the Geometry Slice.
Attributes
----------
nominal_length_mm : int, float
The nominal distance between the geometric nodes, in mm.
"""
nominal_length_mm: float | int = 50
def __init__(
self,
geo_roi1: Point,
geo_roi2: Point,
mm_per_pixel: float,
tolerance: int | float,
):
"""
Parameters
----------
geo_roi1 : GEO_ROI
One of two ROIs representing one end of the line.
geo_roi2 : GEO_ROI
The other ROI which is the other end of the line.
mm_per_pixel : float
The mm/pixel value.
tolerance : int, float
The tolerance of the geometric line, in mm.
"""
super().__init__(geo_roi1, geo_roi2)
self.mm_per_pixel = mm_per_pixel
self.tolerance = tolerance
@property
def passed(self) -> bool:
"""Whether the line passed tolerance."""
return (
self.nominal_length_mm - self.tolerance
< self.length_mm
< self.nominal_length_mm + self.tolerance
)
@property
def pass_fail_color(self) -> str:
"""Plot color for the line, based on pass/fail status."""
return "blue" if self.passed else "red"
@property
def length_mm(self) -> float:
"""Return the length of the line in mm."""
return self.length * self.mm_per_pixel
[docs]
class CTP515(CatPhanModule):
"""Class for analysis of the low contrast slice of the CTP module. Low contrast is measured by obtaining
the average pixel value of the contrast ROIs and comparing that value to the average background value. To obtain
a more "human" detection level, the contrast (which is largely the same across different-sized ROIs) is multiplied
by the diameter. This value is compared to the contrast threshold to decide if it can be "seen".
"""
attr_name = "ctp515"
common_name = "Low Contrast"
num_slices = 1
roi_dist_mm = 50
roi_radius_mm = [6, 3.5, 3, 2.5, 2, 1.5]
roi_angles = [-87.4, -69.1, -52.7, -38.5, -25.1, -12.9]
roi_settings = {
"15": {
"angle": roi_angles[0],
"distance": roi_dist_mm,
"radius": roi_radius_mm[0],
},
"9": {
"angle": roi_angles[1],
"distance": roi_dist_mm,
"radius": roi_radius_mm[1],
},
"8": {
"angle": roi_angles[2],
"distance": roi_dist_mm,
"radius": roi_radius_mm[2],
},
"7": {
"angle": roi_angles[3],
"distance": roi_dist_mm,
"radius": roi_radius_mm[3],
},
"6": {
"angle": roi_angles[4],
"distance": roi_dist_mm,
"radius": roi_radius_mm[4],
},
"5": {
"angle": roi_angles[5],
"distance": roi_dist_mm,
"radius": roi_radius_mm[5],
},
}
background_roi_dist_ratio = 0.75
background_roi_radius_mm = 4
WINDOW_SIZE = 50
def __init__(
self,
catphan,
tolerance: float,
cnr_threshold: float,
offset: int,
contrast_method: str,
visibility_threshold: float,
clear_borders: bool = True,
):
self.cnr_threshold = cnr_threshold
self.contrast_method = contrast_method
self.visibility_threshold = visibility_threshold
super().__init__(
catphan, tolerance=tolerance, offset=offset, clear_borders=clear_borders
)
def _setup_rois(self):
# create both background rois dynamically, then create the actual sample ROI as normal
for name, setting in self.roi_settings.items():
self.background_rois[name + "-outer"] = LowContrastDiskROI(
self.image,
setting["angle_corrected"],
self.background_roi_radius_mm / self.mm_per_pixel,
setting["distance_pixels"] * (2 - self.background_roi_dist_ratio),
self.phan_center,
)
self.background_rois[name + "-inner"] = LowContrastDiskROI(
self.image,
setting["angle_corrected"],
self.background_roi_radius_mm / self.mm_per_pixel,
setting["distance_pixels"] * self.background_roi_dist_ratio,
self.phan_center,
)
background_val = float(
np.mean(
[
self.background_rois[name + "-outer"].pixel_value,
self.background_rois[name + "-inner"].pixel_value,
]
)
)
self.rois[name] = LowContrastDiskROI(
self.image,
setting["angle_corrected"],
setting["radius_pixels"],
setting["distance_pixels"],
self.phan_center,
contrast_reference=background_val,
cnr_threshold=self.cnr_threshold,
contrast_method=self.contrast_method,
visibility_threshold=self.visibility_threshold,
)
@property
def rois_visible(self) -> int:
"""The number of ROIs "visible"."""
return sum(roi.passed_visibility for roi in self.rois.values())
@property
def window_min(self) -> float:
"""Lower bound of CT window/leveling to show on the plotted image. Improves apparent contrast."""
return (
Enumerable(self.background_rois.values()).min(lambda r: r.pixel_value)
- self.WINDOW_SIZE
)
@property
def window_max(self) -> float:
"""Upper bound of CT window/leveling to show on the plotted image. Improves apparent contrast"""
return (
Enumerable(self.rois.values()).max(lambda r: r.pixel_value)
+ self.WINDOW_SIZE
)
class CTP515CP600(CTP515):
roi_angles = [
-87.4 + 180,
-69.1 + 180,
-52.7 + 180,
-38.5 + 180,
-25.1 + 180,
-12.9 + 180,
]
roi_dist_mm = 50
roi_radius_mm = [6, 3.5, 3, 2.5, 2, 1.5]
roi_settings = {
"15": {
"angle": roi_angles[0],
"distance": roi_dist_mm,
"radius": roi_radius_mm[0],
},
"9": {
"angle": roi_angles[1],
"distance": roi_dist_mm,
"radius": roi_radius_mm[1],
},
"8": {
"angle": roi_angles[2],
"distance": roi_dist_mm,
"radius": roi_radius_mm[2],
},
"7": {
"angle": roi_angles[3],
"distance": roi_dist_mm,
"radius": roi_radius_mm[3],
},
"6": {
"angle": roi_angles[4],
"distance": roi_dist_mm,
"radius": roi_radius_mm[4],
},
"5": {
"angle": roi_angles[5],
"distance": roi_dist_mm,
"radius": roi_radius_mm[5],
},
}
class CatPhanBase(ResultsDataMixin[CatphanResult]):
"""A class for loading and analyzing CT DICOM files of a CatPhan 504 & CatPhan 503. Can be from a CBCT or CT scanner
Analyzes: Uniformity (CTP486), High-Contrast Spatial Resolution (CTP528), Image Scaling & HU Linearity (CTP404).
"""
_demo_url: str = ""
_model: str = ""
air_bubble_radius_mm: int | float = 7
localization_radius: int | float = 59
was_from_zip: bool = False
min_num_images = 39
clear_borders: bool = True
hu_origin_slice_variance = 400 # the HU variance required on the origin slice
_phantom_center_func: tuple[Callable, Callable] | None = None
modules: dict[CatPhanModule, dict[str, int]]
dicom_stack: image.DicomImageStack | image.LazyDicomImageStack
def __init__(
self,
folderpath: str | Sequence[str] | Path | Sequence[Path] | Sequence[BytesIO],
check_uid: bool = True,
memory_efficient_mode: bool = False,
):
"""
Parameters
----------
folderpath : str, list of strings, or Path to folder
String that points to the CBCT image folder location.
check_uid : bool
Whether to enforce raising an error if more than one UID is found in the dataset.
memory_efficient_mode : bool
Whether to use a memory efficient mode. If True, the DICOM stack will be loaded on demand rather than all at once.
This will reduce the memory footprint but will be slower by ~25%. Default is False.
Raises
------
NotADirectoryError
If folder str passed is not a valid directory.
FileNotFoundError
If no CT images are found in the folder
"""
self.origin_slice = 0
self.catphan_roll = 0
if isinstance(folderpath, (str, Path)):
if not osp.isdir(folderpath):
raise NotADirectoryError("Path given was not a Directory/Folder")
stack = (
image.DicomImageStack
if not memory_efficient_mode
else image.LazyDicomImageStack
)
self.dicom_stack = stack(
folderpath, check_uid=check_uid, min_number=self.min_num_images
)
@classmethod
def from_demo_images(cls):
"""Construct a CBCT object from the demo images."""
demo_file = retrieve_demo_file(name=cls._demo_url)
return cls.from_zip(demo_file)
@classmethod
def from_url(cls, url: str, check_uid: bool = True):
"""Instantiate a CBCT object from a URL pointing to a .zip object.
Parameters
----------
url : str
URL pointing to a zip archive of CBCT images.
check_uid : bool
Whether to enforce raising an error if more than one UID is found in the dataset.
"""
filename = get_url(url)
return cls.from_zip(filename, check_uid=check_uid)
@classmethod
def from_zip(
cls,
zip_file: str | zipfile.ZipFile | BinaryIO,
check_uid: bool = True,
memory_efficient_mode: bool = False,
):
"""Construct a CBCT object and pass the zip file.
Parameters
----------
zip_file : str, ZipFile
Path to the zip file or a ZipFile object.
check_uid : bool
Whether to enforce raising an error if more than one UID is found in the dataset.
memory_efficient_mode : bool
Whether to use a memory efficient mode. If True, the DICOM stack will be loaded on demand rather than all at once.
This will reduce the memory footprint but will be slower by ~25%. Default is False.
Raises
------
FileExistsError : If zip_file passed was not a legitimate zip file.
FileNotFoundError : If no CT images are found in the folder
"""
delete = not memory_efficient_mode
with TemporaryZipDirectory(zip_file, delete=delete) as temp_zip:
obj = cls(
temp_zip,
check_uid=check_uid,
memory_efficient_mode=memory_efficient_mode,
)
obj.was_from_zip = True
return obj
def plot_analyzed_image(self, show: bool = True, **plt_kwargs) -> None:
"""Plot the images used in the calculation and summary data.
Parameters
----------
show : bool
Whether to plot the image or not.
plt_kwargs : dict
Keyword args passed to the plt.figure() method. Allows one to set things like figure size.
"""
# set up grid and axes
plt.figure(**plt_kwargs)
grid_size = (2, 4)
hu_ax = plt.subplot2grid(grid_size, (0, 1))
self.ctp404.plot(hu_ax)
hu_lin_ax = plt.subplot2grid(grid_size, (0, 2))
self.ctp404.plot_linearity(hu_lin_ax)
# plot side view w/ module locations
side_ax = plt.subplot2grid(grid_size, (1, 2))
self.plot_side_view(side_ax)
# plot individual modules
if self._has_module(CTP486):
unif_ax = plt.subplot2grid(grid_size, (0, 0))
self.ctp486.plot(unif_ax)
unif_prof_ax = plt.subplot2grid(grid_size, (1, 3))
self.ctp486.plot_profiles(unif_prof_ax)
if self._has_module(CTP528CP504):
sr_ax = plt.subplot2grid(grid_size, (1, 0))
self.ctp528.plot(sr_ax)
mtf_ax = plt.subplot2grid(grid_size, (0, 3))
self.ctp528.mtf.plot(mtf_ax)
if self._has_module(CTP515):
locon_ax = plt.subplot2grid(grid_size, (1, 1))
self.ctp515.plot(locon_ax)
# finish up
plt.tight_layout()
if show:
plt.show()
def save_analyzed_image(self, filename: str | Path | BinaryIO, **kwargs) -> None:
"""Save the analyzed summary plot.
Parameters
----------
filename : str, file object
The name of the file to save the image to.
kwargs :
Any valid matplotlib kwargs.
"""
self.plot_analyzed_image(show=False)
plt.savefig(filename, **kwargs)
def plot_analyzed_subimage(
self,
subimage: str = "hu",
delta: bool = True,
show: bool = True,
) -> plt.Figure | None:
"""Plot a specific component of the CBCT analysis.
Parameters
----------
subimage : {'hu', 'un', 'sp', 'lc', 'mtf', 'lin', 'prof', 'side'}
The subcomponent to plot. Values must contain one of the following letter combinations.
E.g. ``linearity``, ``linear``, and ``lin`` will all draw the HU linearity values.
* ``hu`` draws the HU linearity image.
* ``un`` draws the HU uniformity image.
* ``sp`` draws the Spatial Resolution image.
* ``lc`` draws the Low Contrast image (if applicable).
* ``mtf`` draws the RMTF plot.
* ``lin`` draws the HU linearity values. Used with ``delta``.
* ``prof`` draws the HU uniformity profiles.
* ``side`` draws the side view of the phantom with lines of the module locations.
delta : bool
Only for use with ``lin``. Whether to plot the HU delta or actual values.
show : bool
Whether to actually show the plot.
"""
subimage = subimage.lower()
fig, ax = plt.subplots()
plt.axis("off")
if "hu" in subimage: # HU, GEO & thickness objects
self.ctp404.plot(ax)
plt.autoscale(tight=True)
elif "un" in subimage: # uniformity
self.ctp486.plot(ax)
plt.autoscale(tight=True)
elif "sp" in subimage: # SR objects
self.ctp528.plot(ax)
plt.autoscale(tight=True)
elif "mtf" in subimage:
plt.axis("on")
self.ctp528.mtf.plot(ax)
elif "lc" in subimage:
if self._has_module(CTP515):
self.ctp515.plot(ax)
plt.autoscale(tight=True)
else:
return
elif "lin" in subimage:
plt.axis("on")
self.ctp404.plot_linearity(ax, delta)
elif "prof" in subimage:
plt.axis("on")
self.ctp486.plot_profiles(ax)
elif "side" in subimage:
ax = plt.gca()
self.plot_side_view(ax)
else:
raise ValueError(f"Subimage parameter {subimage} not understood")
if show:
plt.show()
return fig
def save_analyzed_subimage(
self,
filename: str | BinaryIO,
subimage: str = "hu",
delta: bool = True,
**kwargs,
) -> plt.Figure | None:
"""Save a component image to file.
Parameters
----------
filename : str, file object
The file to write the image to.
subimage : str
See :meth:`~pylinac.ct.CatPhanBase.plot_analyzed_subimage` for parameter info.
delta : bool
Only for use with ``lin``. Whether to plot the HU delta or actual values.
"""
fig = self.plot_analyzed_subimage(subimage, delta=delta, show=False)
if fig: # no fig if we plot low contrast
plt.savefig(filename, **kwargs)
if isinstance(filename, str):
print(f"CatPhan subimage figure saved to {osp.abspath(filename)}")
return fig
def _results(self) -> None:
"""Helper function to spit out values that will be tested."""
print(self.results())
print(f"Phantom roll: {self.catphan_roll}")
print(f"Origin slice: {self.origin_slice}")
mtfs = {}
for mtf in (95, 90, 80, 50, 30):
mtfval = self.ctp528.mtf.relative_resolution(mtf)
mtfs[mtf] = mtfval
print(f"MTFs: {mtfs}")
def localize(self) -> None:
"""Find the slice number of the catphan's HU linearity module and roll angle"""
self._phantom_center_func = self.find_phantom_axis()
self.origin_slice = self.find_origin_slice()
self.catphan_roll = self.find_phantom_roll()
self.origin_slice = self.refine_origin_slice(
initial_slice_num=self.origin_slice
)
# now that we have the origin slice, ensure we have scanned all linked modules
if not self._ensure_physical_scan_extent():
raise ValueError(
"The physical scan extent does not match the module configuration. "
"This means not all modules were included in the scan. Rescan the phantom to include all"
"relevant modules, or remove modules from the analysis."
)
def _module_offsets(self) -> list[float]:
"""A list of the module offsets. Used to confirm scan extent"""
absolute_origin_position = self.dicom_stack[self.origin_slice].z_position
return [
absolute_origin_position + config["offset"]
for config in self.modules.values()
]
def _ensure_physical_scan_extent(self) -> bool:
"""Ensure that all the modules of the phantom have been scanned. If a CBCT isn't
positioned correctly, some modules might not be included.
It appears there can be rounding errors between the DICOM tag and the actual slice position. See RAM-2897.
"""
z_positions = [z_position(m) for m in self.dicom_stack.metadatas]
min_scan_extent_slice = round(min(z_positions), 1)
max_scan_extent_slice = round(max(z_positions), 1)
min_config_extent_slice = round(min(self._module_offsets()), 1)
max_config_extent_slice = round(max(self._module_offsets()), 1)
return (min_config_extent_slice >= min_scan_extent_slice) and (
max_config_extent_slice <= max_scan_extent_slice
)
def find_phantom_axis(self) -> (Callable, Callable):
"""We fit all the center locations of the phantom across all slices to a 1D poly function instead of finding them individually for robustness.
Normally, each slice would be evaluated individually, but the RadMachine jig gets in the way of
detecting the HU module (🤦♂️). To work around that in a backwards-compatible way we instead
look at all the slices and if the phantom was detected, capture the phantom center.
ALL the centers are then fitted to a 1D poly function and passed to the individual slices.
This way, even if one slice is messed up (such as because of the phantom jig), the poly function
is robust to give the real center based on all the other properly-located positions on the other slices.
"""
z = []
center_x = []
center_y = []
for idx, img in enumerate(self.dicom_stack):
slice = Slice(
self,
slice_num=idx,
clear_borders=self.clear_borders,
original_image=img,
)
if slice.is_phantom_in_view():
roi = slice.phantom_roi
z.append(idx)
center_y.append(roi.centroid[0])
center_x.append(roi.centroid[1])
# clip to exclude any crazy values
zs = np.array(z)
center_xs = np.array(center_x)
center_ys = np.array(center_y)
# gives an absolute and relative range so tight ranges are all included
# but extreme values are excluded. Sometimes the range is very tight
# and thus percentiles are not a sure thing
x_idxs = np.argwhere(
np.isclose(np.median(center_xs), center_xs, atol=3, rtol=0.01)
)
y_idxs = np.argwhere(
np.isclose(np.median(center_ys), center_ys, atol=3, rtol=0.01)
)
common_idxs = np.intersect1d(x_idxs, y_idxs)
# fit to 1D polynomials; inspiration: https://stackoverflow.com/a/45351484
fit_zx = np.poly1d(np.polyfit(zs[common_idxs], center_xs[common_idxs], deg=1))
fit_zy = np.poly1d(np.polyfit(zs[common_idxs], center_ys[common_idxs], deg=1))
return fit_zx, fit_zy
@property
def mm_per_pixel(self) -> float:
"""The millimeters per pixel of the DICOM images."""
return self.dicom_stack.metadata.PixelSpacing[0]
def find_origin_slice(self) -> int:
"""Using a brute force search of the images, find the median HU linearity slice.
This method walks through all the images and takes a collapsed circle profile where the HU
linearity ROIs are. If the profile contains both low (<800) and high (>800) HU values and most values are the same
(i.e. it's not an artifact), then
it can be assumed it is an HU linearity slice. The median of all applicable slices is the
center of the HU slice.
Returns
-------
int
The middle slice of the HU linearity module.
"""
hu_slices = []
for image_number in range(0, self.num_images, 2):
slice = Slice(
self, image_number, combine=False, clear_borders=self.clear_borders
)
# print(image_number)
# slice.image.plot()
if slice.is_phantom_in_view():
circle_prof = CollapsedCircleProfile(
slice.phan_center,
radius=self.localization_radius / self.mm_per_pixel,
image_array=slice.image,
width_ratio=0.05,
num_profiles=5,
)
prof = circle_prof.values
# determine if the profile contains both low and high values and that most values are the same
low_end, high_end = np.percentile(prof, [2, 98])
median = np.median(prof)
middle_variation = np.percentile(prof, 80) - np.percentile(prof, 20)
variation_limit = max(
100, self.dicom_stack.metadata.SliceThickness * -100 + 300
)
if (
(low_end < median - self.hu_origin_slice_variance)
and (high_end > median + self.hu_origin_slice_variance)
and (middle_variation < variation_limit)
):
hu_slices.append(image_number)
if not hu_slices:
raise ValueError(
"No slices were found that resembled the HU linearity module"
)
hu_slices = np.array(hu_slices)
c = int(round(float(np.median(hu_slices))))
ln = len(hu_slices)
# drop slices that are way far from median
hu_slices = hu_slices[((c + ln / 2) >= hu_slices) & (hu_slices >= (c - ln / 2))]
center_hu_slice = int(round(float(np.median(hu_slices))))
if self._is_within_image_extent(center_hu_slice):
return center_hu_slice
def refine_origin_slice(self, initial_slice_num: int) -> int:
"""Apply a refinement to the origin slice. This was added to handle
the catphan 604 at least due to variations in the length of the HU plugs."""
return initial_slice_num
def _is_right_area(self, region: RegionProperties):
thresh = np.pi * ((self.air_bubble_radius_mm / self.mm_per_pixel) ** 2)
return thresh * 2 > region.filled_area > thresh / 2
def _is_right_eccentricity(self, region: RegionProperties):
return region.eccentricity < 0.5
def find_phantom_roll(self, func: Callable | None = None) -> float:
"""Determine the "roll" of the phantom.
This algorithm uses the two air bubbles in the HU slice and the resulting angle between them.
Parameters
----------
func
A callable to sort the air ROIs.
Returns
-------
float : the angle of the phantom in **degrees**.
"""
# get edges and make ROIs from it
slice = Slice(self, self.origin_slice, clear_borders=self.clear_borders)
larr, regions, _ = get_regions(slice)
# find appropriate ROIs and grab the two most centrally positioned ones
hu_bubbles = [
r
for r in regions
if (self._is_right_area(r) and self._is_right_eccentricity(r))
]
func = func or (lambda x: abs(x.centroid[1] - slice.phan_center.x))
central_bubbles = sorted(hu_bubbles, key=func)[:2]
sorted_bubbles = sorted(
central_bubbles, key=lambda x: x.centroid[0]
) # top, bottom
y_dist = sorted_bubbles[1].centroid[0] - sorted_bubbles[0].centroid[0]
x_dist = sorted_bubbles[1].centroid[1] - sorted_bubbles[0].centroid[1]
phan_roll = np.arctan2(y_dist, x_dist)
anglroll = np.rad2deg(phan_roll) - 90
return anglroll
@property
def num_images(self) -> int:
"""The number of images loaded."""
return len(self.dicom_stack)
def _is_within_image_extent(self, image_num: int) -> bool:
"""Determine if the image number is beyond the edges of the images (negative or past last image)."""
if self.num_images - 1 > image_num > 1:
return True
else:
raise ValueError(
"The determined image number is beyond the image extent. Either the entire dataset "
"wasn't loaded or the entire phantom wasn't scanned."
)
@property
def catphan_size(self) -> float:
"""The expected size of the phantom in pixels, based on a 20cm wide phantom."""
phan_area = np.pi * (self.catphan_radius_mm**2)
return phan_area / (self.mm_per_pixel**2)
def publish_pdf(
self,
filename: str | Path,
notes: str | None = None,
open_file: bool = False,
metadata: dict | None = None,
logo: Path | str | None = None,
) -> None:
"""Publish (print) a PDF containing the analysis and quantitative results.
Parameters
----------
filename : (str, file-like object}
The file to write the results to.
notes : str, list of strings
Text; if str, prints single line.
If list of strings, each list item is printed on its own line.
open_file : bool
Whether to open the file using the default program after creation.
metadata : dict
Extra data to be passed and shown in the PDF. The key and value will be shown with a colon.
E.g. passing {'Author': 'James', 'Unit': 'TrueBeam'} would result in text in the PDF like:
--------------
Author: James
Unit: TrueBeam
--------------
logo: Path, str
A custom logo to use in the PDF report. If nothing is passed, the default pylinac logo is used.
"""
analysis_title = f"CatPhan {self._model} Analysis"
module_images = [("hu", "lin")]
if self._has_module(CTP528CP504):
module_images.append(("sp", "mtf"))
if self._has_module(CTP486):
module_images.append(("un", "prof"))
if self._has_module(CTP515):
module_images.append(("lc", None))
module_images.append(("side", None))
self._publish_pdf(
filename,
metadata,
notes,
analysis_title,
[*self.results(as_list=True), ""],
module_images,
logo,
)
if open_file:
webbrowser.open(filename)
def _publish_pdf(
self,
filename: str,
metadata: dict | None,
notes: str,
analysis_title: str,
texts: Sequence[str],
imgs: Sequence[tuple[str, str]],
logo: Path | str | None = None,
):
canvas = pdf.PylinacCanvas(
filename, page_title=analysis_title, metadata=metadata, logo=logo
)
if notes is not None:
canvas.add_text(text="Notes:", location=(1, 4.5), font_size=14)
canvas.add_text(text=notes, location=(1, 4))
for page, ((img1, img2), text) in enumerate(zip(imgs, texts)):
for img, offset in zip((img1, img2), (12, 2)):
if img is not None:
data = io.BytesIO()
self.save_analyzed_subimage(data, img)
canvas.add_image(data, location=(4, offset), dimensions=(15, 10))
canvas.add_text(text=text, location=(1.5, 23))
canvas.add_new_page()
canvas.finish()
def _zip_images(self) -> None:
"""Compress the raw images into a ZIP archive and remove the uncompressed images."""
zip_name = rf'{osp.dirname(self.dicom_stack[0].path)}\CBCT - {self.dicom_stack[0].date_created(format="%A, %I-%M-%S, %B %d, %Y")}.zip'
with zipfile.ZipFile(zip_name, "w", compression=zipfile.ZIP_DEFLATED) as zfile:
for img in self.dicom_stack:
zfile.write(img.path, arcname=osp.basename(img.path))
for img in self.dicom_stack:
try:
os.remove(img.path)
except Exception:
pass
def plot_side_view(self, axis: Axes) -> None:
"""Plot a view of the scan from the side with lines showing detected module positions"""
side_array = self.dicom_stack.side_view(axis=1)
axis.set_yticks([])
axis.set_title("Side View")
axis.imshow(side_array, aspect="auto", cmap="gray", interpolation="none")
for module in self._detected_modules():
axis.axvline(module.slice_num)
def _detected_modules(self) -> list[CatPhanModule]:
"""A list of the modules detected. Unlike _get_module, this returns the instances"""
modules = [self.ctp404]
if self._has_module(CTP515):
modules.append(self.ctp515)
if self._has_module(CTP486):
modules.append(self.ctp486)
if self._has_module(CTP528CP504):
modules.append(self.ctp528)
return modules
def analyze(
self,
hu_tolerance: int | float = 40,
scaling_tolerance: int | float = 1,
thickness_tolerance: int | float = 0.2,
low_contrast_tolerance: int | float = 1,
cnr_threshold: int | float = 15,
zip_after: bool = False,
contrast_method: str = Contrast.MICHELSON,
visibility_threshold: float = 0.15,
thickness_slice_straddle: str | int = "auto",
expected_hu_values: dict[str, int | float] | None = None,
):
"""Single-method full analysis of CBCT DICOM files.
Parameters
----------
hu_tolerance : int
The HU tolerance value for both HU uniformity and linearity.
scaling_tolerance : float, int
The scaling tolerance in mm of the geometric nodes on the HU linearity slice (CTP404 module).
thickness_tolerance : float, int
The tolerance of the thickness calculation in mm, based on the wire ramps in the CTP404 module.
.. warning:: Thickness accuracy degrades with image noise; i.e. low mAs images are less accurate.
low_contrast_tolerance : int
The number of low-contrast bubbles needed to be "seen" to pass.
cnr_threshold : float, int
The threshold for "detecting" low-contrast image. See RTD for calculation info.
.. deprecated:: 3.0
Use visibility parameter instead.
zip_after : bool
If the CT images were not compressed before analysis and this is set to true, pylinac will compress
the analyzed images into a ZIP archive.
contrast_method
The contrast equation to use. See :ref:`low_contrast_topic`.
visibility_threshold
The threshold for detecting low-contrast ROIs. Use instead of ``cnr_threshold``. Follows the Rose equation.
See :ref:`visibility`.
thickness_slice_straddle
The number of extra slices **on each side** of the HU module slice to use for slice thickness determination.
The rationale is that for thin slices the ramp FWHM can be very noisy. I.e. a 1mm slice might have a 100%
variation with a low-mAs protocol. To account for this, slice thicknesses < 3.5mm have 1 slice added
on either side of the HU module (so 3 total slices) and then averaged. The default is 'auto',
which follows the above logic. Set to an integer to explicitly use a certain amount of padding. Typical
values are 0, 1, and 2.
.. warning:: This is the padding **on either side**. So a value of 1 => 3 slices, 2 => 5 slices, 3 => 7 slices, etc.
expected_hu_values
An optional dictionary of the expected HU values for the HU linearity module. The keys are the ROI names and the values
are the expected HU values. If a key is not present or the parameter is None, the default values will be used.
"""
self.localize()
ctp404, offset = self._get_module(CTP404CP504, raise_empty=True)
self.ctp404 = ctp404(
self,
offset=offset,
hu_tolerance=hu_tolerance,
thickness_tolerance=thickness_tolerance,
scaling_tolerance=scaling_tolerance,
clear_borders=self.clear_borders,
thickness_slice_straddle=thickness_slice_straddle,
expected_hu_values=expected_hu_values,
)
if self._has_module(CTP486):
ctp486, offset = self._get_module(CTP486)
self.ctp486 = ctp486(
self,
offset=offset,
tolerance=hu_tolerance,
clear_borders=self.clear_borders,
)
if self._has_module(CTP528CP504):
ctp528, offset = self._get_module(CTP528CP504)
self.ctp528 = ctp528(
self, offset=offset, tolerance=None, clear_borders=self.clear_borders
)
if self._has_module(CTP515):
ctp515, offset = self._get_module(CTP515)
self.ctp515 = ctp515(
self,
tolerance=low_contrast_tolerance,
cnr_threshold=cnr_threshold,
offset=offset,
contrast_method=contrast_method,
visibility_threshold=visibility_threshold,
clear_borders=self.clear_borders,
)
if zip_after and not self.was_from_zip:
self._zip_images()
def _has_module(self, module_of_interest: type[CatPhanModule]) -> bool:
return any(
issubclass(module, module_of_interest) for module in self.modules.keys()
)
def _get_module(
self, module_of_interest: type[CatPhanModule], raise_empty: bool = False
) -> tuple[type[CatPhanModule], int]:
"""Grab the module that is, or is a subclass of, the module of interest. This allows users to subclass a CTP module and pass that in."""
for module, values in self.modules.items():
if issubclass(module, module_of_interest):
return module, values.get("offset")
if raise_empty:
raise ValueError(
f"Tried to find the {module_of_interest} or a subclass of it. Did you override `modules` and not pass this module in?"
)
def results(self, as_list: bool = False) -> str | list[list[str]]:
"""Return the results of the analysis as a string. Use with print().
Parameters
----------
as_list : bool
Whether to return as a list of list of strings vs single string. Pretty much for internal usage.
"""
results = []
result = [
f" - CBCT/CT {self._model} QA Test - ",
" - CTP 404 Results - ",
f"HU Linearity tolerance: {self.ctp404.hu_tolerance}",
"HU Linearity ROIs:",
# wrap so it doesn't fall off the page in PDFs
*textwrap.wrap(self.ctp404.roi_vals_as_str, width=50),
f"HU Passed?: {self.ctp404.passed_hu}",
f"Low contrast visibility: {self.ctp404.lcv:2.2f}",
f"Geometric Line Average (mm): {self.ctp404.avg_line_length:2.2f}",
f"Geometry Passed?: {self.ctp404.passed_geometry}",
f"Measured Slice Thickness (mm): {self.ctp404.meas_slice_thickness:2.3f}",
f"Slice Thickness Passed? {self.ctp404.passed_thickness}",
]
results.append(result)
if self._has_module(CTP528CP504):
ctp528_result = [
" - CTP528 Results - ",
f"MTF 80% (lp/mm): {self.ctp528.mtf.relative_resolution(80):2.2f}",
f"MTF 50% (lp/mm): {self.ctp528.mtf.relative_resolution(50):2.2f}",
f"MTF 30% (lp/mm): {self.ctp528.mtf.relative_resolution(30):2.2f}",
]
results.append(ctp528_result)
if self._has_module(CTP486):
ctp486_result = [
" - CTP486 Results - ",
f"Uniformity tolerance: {self.ctp486.tolerance}",
f"Uniformity ROIs: {self.ctp486.roi_vals_as_str}",
f"Uniformity index: {self.ctp486.uniformity_index:2.3f}",
f"Integral non-uniformity: {self.ctp486.integral_non_uniformity:2.4f}",
f"Uniformity Passed?: {self.ctp486.overall_passed}",
f"Max Noise Power frequency: {self.ctp486.max_noise_power_frequency}",
f"Average Noise Power: {self.ctp486.avg_noise_power}",
]
results.append(ctp486_result)
if self._has_module(CTP515):
ctp515_result = [
" - CTP515 Results - ",
f"CNR threshold: {self.ctp515.cnr_threshold}",
f'Low contrast ROIs "seen": {self.ctp515.rois_visible}',
]
results.append(ctp515_result)
if not as_list:
result = "\n".join(itertools.chain(*results))
else:
result = results
return result
def _generate_results_data(self) -> CatphanResult:
"""Present the results data and metadata as a dataclass or dict.
The default return type is a dataclass."""
ctp404_result = CTP404Result(
offset=self.ctp404._offset,
low_contrast_visibility=self.ctp404.lcv,
thickness_passed=self.ctp404.passed_thickness,
measured_slice_thickness_mm=self.ctp404.meas_slice_thickness,
thickness_num_slices_combined=self.ctp404.num_slices + self.ctp404.pad,
geometry_passed=self.ctp404.passed_geometry,
avg_line_distance_mm=self.ctp404.avg_line_length,
line_distances_mm=[
line.length_mm for name, line in self.ctp404.lines.items()
],
hu_linearity_passed=self.ctp404.passed_hu,
hu_tolerance=self.ctp404.hu_tolerance,
hu_rois=rois_to_results(self.ctp404.rois),
)
data = CatphanResult(
catphan_model=self._model,
catphan_roll_deg=self.catphan_roll,
origin_slice=self.origin_slice,
num_images=self.num_images,
ctp404=ctp404_result,
)
# CTP 486 Uniformity stuff
if self._has_module(CTP486):
data.ctp486 = CTP486Result(
passed=self.ctp486.overall_passed,
uniformity_index=self.ctp486.uniformity_index,
integral_non_uniformity=self.ctp486.integral_non_uniformity,
rois=rois_to_results(self.ctp486.rois),
nps_avg_power=self.ctp486.avg_noise_power,
nps_max_freq=self.ctp486.max_noise_power_frequency,
)
# CTP 528 stuff
if self._has_module(CTP528CP504):
data.ctp528 = CTP528Result(
roi_settings=self.ctp528.roi_settings,
start_angle_radians=self.ctp528.start_angle,
mtf_lp_mm={
p: self.ctp528.mtf.relative_resolution(p) for p in range(10, 91, 10)
},
)
# CTP 515 stuff
if self._has_module(CTP515):
data.ctp515 = CTP515Result(
cnr_threshold=self.ctp515.cnr_threshold,
num_rois_seen=self.ctp515.rois_visible,
roi_settings=self.ctp515.roi_settings,
roi_results={
key: roi.as_dict() for key, roi in self.ctp515.rois.items()
},
)
return data
[docs]
class CatPhan503(CatPhanBase):
"""A class for loading and analyzing CT DICOM files of a CatPhan 503.
Analyzes: Uniformity (CTP486), High-Contrast Spatial Resolution (CTP528), Image Scaling & HU Linearity (CTP404).
"""
_demo_url = "CatPhan503.zip"
_model = "503"
catphan_radius_mm = 97
modules = {
CTP404CP503: {"offset": 0},
CTP486: {"offset": -110},
CTP528CP503: {"offset": -30},
}
[docs]
@staticmethod
def run_demo(show: bool = True):
"""Run the CBCT demo using high-quality head protocol images."""
cbct = CatPhan503.from_demo_images()
cbct.analyze()
print(cbct.results())
cbct.plot_analyzed_image(show)
[docs]
class CatPhan504(CatPhanBase):
"""A class for loading and analyzing CT DICOM files of a CatPhan 504. Can be from a CBCT or CT scanner
Analyzes: Uniformity (CTP486), High-Contrast Spatial Resolution (CTP528),
Image Scaling & HU Linearity (CTP404), and Low contrast (CTP515).
"""
_demo_url = "CatPhan504.zip"
_model = "504"
catphan_radius_mm = 101
modules = {
CTP404CP504: {"offset": 0},
CTP486: {"offset": -65},
CTP528CP504: {"offset": 30},
CTP515: {"offset": -30},
}
[docs]
@staticmethod
def run_demo(show: bool = True):
"""Run the CBCT demo using high-quality head protocol images."""
cbct = CatPhan504.from_demo_images()
cbct.analyze()
print(cbct.results())
cbct.plot_analyzed_image(show)
[docs]
class CatPhan604(CatPhanBase):
"""A class for loading and analyzing CT DICOM files of a CatPhan 604. Can be from a CBCT or CT scanner
Analyzes: Uniformity (CTP486), High-Contrast Spatial Resolution (CTP528),
Image Scaling & HU Linearity (CTP404), and Low contrast (CTP515).
"""
_demo_url = "CatPhan604.zip"
_model = "604"
catphan_radius_mm = 101
modules = {
CTP404CP604: {"offset": 0},
CTP486: {"offset": -80},
CTP528CP604: {"offset": 40},
CTP515: {"offset": -40},
}
[docs]
@staticmethod
def run_demo(show: bool = True):
"""Run the CBCT demo using high-quality head protocol images."""
cbct = CatPhan604.from_demo_images()
cbct.analyze()
print(cbct.results())
cbct.plot_analyzed_image(show)
[docs]
def refine_origin_slice(self, initial_slice_num: int) -> int:
"""The HU plugs are longer than the 'wire section'. This applies a refinement to find the
slice that has the least angle between the centers of the left and right wires.
Starting with the initial slice, go +/- 5 slices to find the slice with the least angle
between the left and right wires.
This suffers from a weakness that the roll is not yet determined.
This will thus return the slice that has the least ABSOLUTE
roll. If the phantom has an inherent roll, this will not be detected and may be off by a slice or so.
Given the angle of the wire, the error would be small and likely only 1-2 slices max.
"""
angles = []
# make a CTP module for the purpose of easily extracting the thickness ROIs
ctp404, offset = self._get_module(CTP404CP604, raise_empty=True)
# we don't want to set up our other ROIs (they sometimes fail) so we temporarily override the method
original_setup = ctp404._setup_rois
ctp404._setup_rois = lambda x: x
ctp = ctp404(
self,
offset=offset,
clear_borders=self.clear_borders,
hu_tolerance=0,
scaling_tolerance=0,
thickness_tolerance=0,
)
# we have to reset the method after we're done for future calls
ctp404._setup_rois = original_setup
for slice_num in range(initial_slice_num - 5, initial_slice_num + 5):
slice = Slice(self, slice_num, clear_borders=self.clear_borders)
# make a slice and add the wire ROIs to it.
troi = {}
for name, setting in ctp.thickness_roi_settings.items():
troi[name] = ThicknessROI(
slice.image,
setting["width_pixels"],
setting["height_pixels"],
setting["angle_corrected"],
setting["distance_pixels"],
slice.phan_center,
)
# now find the angle between the left and right and top and bottom wires via the long profile
left_wire = troi["Left"].long_profile.center_idx
right_wire = troi["Right"].long_profile.center_idx
h_angle = abs(left_wire - right_wire)
top_wire = troi["Top"].long_profile.center_idx
bottom_wire = troi["Bottom"].long_profile.center_idx
v_angle = abs(top_wire - bottom_wire)
angle = (h_angle + v_angle) / 2
angles.append(
{
"slice": slice_num,
"angle": angle,
"left width": troi["Left"].long_profile.field_width_px,
"right width": troi["Right"].long_profile.field_width_px,
}
)
# some slices might not include the wire
# we need to drop those; we do so by dropping pairs that have a field width well below the median
median_width_l = np.median([angle["left width"] for angle in angles])
median_width_r = np.median([angle["right width"] for angle in angles])
median_width = (median_width_l + median_width_r) / 2
for angle_set in angles.copy():
if (
angle_set["left width"] < median_width * 0.7
or angle_set["right width"] < median_width * 0.7
):
angles.remove(angle_set)
# now find the slice with the least angle, accounting for the phantom roll
m_slice_num = np.argsort([a["angle"] - self.catphan_roll for a in angles])
refined_slice = angles[m_slice_num[0]]["slice"]
return refined_slice
[docs]
class CatPhan600(CatPhanBase):
"""A class for loading and analyzing CT DICOM files of a CatPhan 600.
Analyzes: Uniformity (CTP486), High-Contrast Spatial Resolution (CTP528),
Image Scaling & HU Linearity (CTP404), and Low contrast (CTP515).
"""
_demo_url = "CatPhan600.zip"
_model = "600"
catphan_radius_mm = 101
modules = {
CTP404CP600: {"offset": 0},
CTP486: {"offset": -160},
CTP515CP600: {"offset": -110},
CTP528CP600: {"offset": -70},
}
[docs]
@staticmethod
def run_demo(show: bool = True):
"""Run the CatPhan 600 demo."""
cbct = CatPhan600.from_demo_images()
cbct.analyze()
print(cbct.results())
cbct.plot_analyzed_image(show)
[docs]
def find_phantom_roll(self, func: Callable | None = None) -> float:
"""With the CatPhan 600, we have to consider that the top air ROI
has a water vial in it (see pg 12 of the manual). If so, the top air ROI won't be detected.
Rather, the default algorithm will find the bottom air ROI and teflon to the left.
It may also find the top air ROI if the water vial isn't there.
We use the below lambda to select the bottom air and teflon ROIs consistently.
These two ROIs are at 75 degrees from cardinal. We thus offset the default outcome by 75.
"""
angle = super().find_phantom_roll(lambda x: -x.centroid[0])
return angle + 75
[docs]
def get_regions(
slice_or_arr: Slice | np.array,
fill_holes: bool = False,
clear_borders: bool = True,
threshold: str = "otsu",
) -> tuple[np.array, list, int]:
"""Get the skimage regions of a black & white image."""
if threshold == "otsu":
thresmeth = filters.threshold_otsu
elif threshold == "mean":
thresmeth = np.mean
if isinstance(slice_or_arr, Slice):
edges = filters.scharr(slice_or_arr.image.array.astype(float))
center = slice_or_arr.image.center
elif isinstance(slice_or_arr, np.ndarray):
edges = filters.scharr(slice_or_arr.astype(float))
center = (int(edges.shape[1] / 2), int(edges.shape[0] / 2))
edges = filters.gaussian(edges, sigma=1)
if isinstance(slice_or_arr, Slice):
radius = 110 / slice_or_arr.mm_per_pixel
rr, cc = draw.disk(
center=(center.y, center.x), radius=radius, shape=edges.shape
)
thres = thresmeth(edges[rr, cc])
else:
thres = thresmeth(edges)
bw = edges > thres
if clear_borders:
bw = segmentation.clear_border(bw, buffer_size=int(max(bw.shape) / 50))
if fill_holes:
bw = ndimage.binary_fill_holes(bw)
labeled_arr, num_roi = measure.label(bw, return_num=True)
regionprops = measure.regionprops(labeled_arr, edges)
return labeled_arr, regionprops, num_roi
[docs]
def combine_surrounding_slices(
dicomstack: DicomImageStack,
nominal_slice_num: int,
slices_plusminus: int = 1,
mode: str = "mean",
) -> np.array:
"""Return an array that is the combination of a given slice and a number of slices surrounding it.
Parameters
----------
dicomstack : `~pylinac.core.image.DicomImageStack`
The CBCT DICOM stack.
nominal_slice_num : int
The slice of interest (along 3rd dim).
slices_plusminus: int
How many slices plus and minus to combine (also along 3rd dim).
mode : {'mean', 'median', 'max}
Specifies the method of combination.
Returns
-------
combined_array : numpy.array
The combined array of the DICOM stack slices.
"""
slices = range(
nominal_slice_num - slices_plusminus, nominal_slice_num + slices_plusminus + 1
)
arrays = tuple(dicomstack[s].array for s in slices)
array_stack = np.dstack(arrays)
if mode == "mean":
combined_array = np.mean(array_stack, 2)
elif mode == "median":
combined_array = np.median(array_stack, 2)
else:
combined_array = np.max(array_stack, 2)
return combined_array
def rois_to_results(dict_mapping: dict[str, DiskROI]) -> dict[str, ROIResult]:
"""Converts a dict of HUDiskROIs to a dict of ROIResults. This is for dumping to simple data formats for results_data and RadMachine"""
flat_dict = {}
for name, roi in dict_mapping.items():
flat_dict[name] = ROIResult(
name=name,
value=roi.pixel_value,
stdev=roi.std,
difference=getattr(roi, "value_diff", None),
nominal_value=getattr(roi, "nominal_val", None),
passed=getattr(roi, "passed", None),
)
return flat_dict