from __future__ import annotations
import io
import math
import textwrap
import warnings
import webbrowser
from io import BytesIO
from pathlib import Path
import numpy as np
from matplotlib import pyplot as plt
from pydantic import BaseModel
from scipy import ndimage
from .core import pdf
from .core.array_utils import find_nearest_idx
from .core.geometry import Line, Point
from .core.mtf import MTF
from .core.profile import FWXMProfilePhysical
from .core.roi import HighContrastDiskROI, RectangleROI
from .core.utilities import ResultBase, ResultsDataMixin
from .ct import (
CatPhanBase,
CatPhanModule,
Slice,
ThicknessROI,
get_regions,
rois_to_results,
)
# CT
CT_UNIFORMITY_MODULE_OFFSET_MM = 70
CT_SPATIAL_RESOLUTION_MODULE_OFFSET_MM = 100
CT_LOW_CONTRAST_MODULE_OFFSET_MM = 30
# MR
MR_SLICE11_MODULE_OFFSET_MM = 100
MR_GEOMETRIC_DISTORTION_MODULE_OFFSET_MM = 40
MR_UNIFORMITY_MODULE_OFFSET_MM = 60
class CTModule(CatPhanModule):
attr_name = "ct_calibration_module"
roi_dist_mm = 63
roi_radius_mm = 10
roi_settings = {
"Air": {"angle": 45, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Poly": {"angle": 225, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Acrylic": {"angle": 135, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Bone": {"angle": -45, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Water": {"angle": 180, "distance": roi_dist_mm, "radius": roi_radius_mm},
}
window_min = -200
window_max = 200
[docs]
class CTModuleOutput(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
roi_distance_from_center_mm: int
roi_radius_mm: int
roi_settings: dict
rois: dict
class UniformityModule(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 = "uniformity_module"
common_name = "HU Uniformity"
roi_dist_mm = 66
roi_radius_mm = 11
roi_settings = {
"Top": {"angle": -90, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Right": {"angle": 0, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Bottom": {"angle": 90, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Left": {"angle": 180, "distance": roi_dist_mm, "radius": roi_radius_mm},
"Center": {"angle": 0, "distance": 0, "radius": roi_radius_mm},
}
window_min = -50
window_max = 50
class SpatialResolutionModule(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 = "spatial_resolution_module"
common_name = "Spatial Resolution"
rois: dict[str, HighContrastDiskROI]
roi_dist_mm = 70
roi_radius_mm = 6
roi_settings = {
"10oclock": {
"angle": -135,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.4,
},
"9oclock": {
"angle": -180,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.5,
},
"7oclock": {
"angle": 135,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.6,
},
"6oclock": {
"angle": 90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.7,
},
"4oclock": {
"angle": 45,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.8,
},
"3oclock": {
"angle": 0,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 0.9,
},
"2oclock": {
"angle": -45,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 1.0,
},
"12oclock": {
"angle": -90,
"distance": roi_dist_mm,
"radius": roi_radius_mm,
"lp/mm": 1.2,
},
}
def _setup_rois(self) -> None:
for name, setting in self.roi_settings.items():
self.rois[name] = HighContrastDiskROI(
self.image,
setting["angle_corrected"],
setting["radius_pixels"],
setting["distance_pixels"],
self.phan_center,
contrast_threshold=1.0, # fixed to 1 so everything passes. We aren't evaluating pass/fail here
)
@property
def mtf(self) -> MTF:
spacings = [roi["lp/mm"] for roi in self.roi_settings.values()]
return MTF.from_high_contrast_diskset(
spacings=spacings, diskset=list(self.rois.values())
)
def plot_rois(self, axis: plt.Axes) -> None:
"""Plot the ROIs to the axis. Override to set the color"""
for roi, mtf in zip(self.rois.values(), self.mtf.norm_mtfs.values()):
roi.plot2axes(axis, edgecolor="g")
[docs]
class SpatialResolutionModuleOutput(CTModuleOutput):
"""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."""
lpmm_to_rmtf: dict #:
class LowContrastModule(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 = "low_contrast_module"
common_name = "Low Contrast"
roi_dist_mm = 60
roi_radius_mm = 6
nominal_value = 0
roi_settings = {
"ROI": {"angle": -90, "distance": roi_dist_mm, "radius": roi_radius_mm},
}
background_roi_settings = {
"ROI": {"angle": -115, "distance": roi_dist_mm, "radius": roi_radius_mm},
}
window_min = 50
window_max = 150
def cnr(self) -> float:
"""Given in the guidance doc as |A-B|/SD where A is the contrast ROI, B is the background, and SD is stdev of B"""
return (
abs(self.rois["ROI"].pixel_value - self.background_rois["ROI"].pixel_value)
/ self.background_rois["ROI"].std
)
[docs]
class LowContrastModuleOutput(CTModuleOutput):
"""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: float #:
[docs]
class ACRCTResult(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."""
phantom_model: str #:
phantom_roll_deg: float #:
origin_slice: int #:
num_images: int #:
ct_module: CTModuleOutput #:
uniformity_module: UniformityModuleOutput #:
low_contrast_module: LowContrastModuleOutput #:
spatial_resolution_module: SpatialResolutionModuleOutput #:
[docs]
class ACRCT(CatPhanBase, ResultsDataMixin[ACRCTResult]):
_model = "ACR CT 464"
catphan_radius_mm = 100
air_bubble_radius_mm = 14
min_num_images = 4
localization_radius = 70
ct_calibration_module = CTModule
low_contrast_module = LowContrastModule
spatial_resolution_module = SpatialResolutionModule
uniformity_module = UniformityModule
clear_borders = False
def _detected_modules(self) -> list[CatPhanModule]:
return [
self.ct_calibration_module,
self.low_contrast_module,
self.spatial_resolution_module,
self.uniformity_module,
]
[docs]
def plot_analyzed_subimage(self, *args, **kwargs):
raise NotImplementedError("Use `plot_images`")
[docs]
def save_analyzed_subimage(self, *args, **kwargs):
raise NotImplementedError("Use `save_images`")
[docs]
def analyze(self) -> None:
"""Analyze the ACR CT phantom"""
self.localize()
self.ct_calibration_module = self.ct_calibration_module(
self, offset=0, clear_borders=self.clear_borders
)
self.uniformity_module = self.uniformity_module(
self,
offset=CT_UNIFORMITY_MODULE_OFFSET_MM,
clear_borders=self.clear_borders,
)
self.spatial_resolution_module = self.spatial_resolution_module(
self,
offset=CT_SPATIAL_RESOLUTION_MODULE_OFFSET_MM,
clear_borders=self.clear_borders,
)
self.low_contrast_module = self.low_contrast_module(
self,
offset=CT_LOW_CONTRAST_MODULE_OFFSET_MM,
clear_borders=self.clear_borders,
)
[docs]
def plot_analyzed_image(self, show: bool = True, **plt_kwargs) -> plt.Figure:
"""Plot the analyzed image
Parameters
----------
show
Whether to show the image.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
# set up grid and axes
fig = plt.figure(**plt_kwargs)
grid_size = (2, 3)
hu_ax = plt.subplot2grid(grid_size, (0, 0))
self.ct_calibration_module.plot(hu_ax)
unif_ax = plt.subplot2grid(grid_size, (0, 1))
self.uniformity_module.plot(unif_ax)
sr_ax = plt.subplot2grid(grid_size, (0, 2))
self.spatial_resolution_module.plot(sr_ax)
locon_ax = plt.subplot2grid(grid_size, (1, 0))
self.low_contrast_module.plot(locon_ax)
spatial_res_graph = plt.subplot2grid(grid_size, (1, 2))
self.spatial_resolution_module.mtf.plot(spatial_res_graph)
side_ax = plt.subplot2grid(grid_size, (1, 1))
self.plot_side_view(side_ax)
# finish up
plt.tight_layout()
if show:
plt.show()
return fig
[docs]
def save_analyzed_image(self, filename: str | Path | BytesIO, **plt_kwargs) -> None:
"""Save the analyzed image to disk or stream
Parameters
----------
filename
Where to save the image to
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
fig = self.plot_analyzed_image(show=False, **plt_kwargs)
fig.savefig(filename)
[docs]
def plot_images(self, show: bool = True, **plt_kwargs) -> dict[str, plt.Figure]:
"""Plot all the individual images separately
Parameters
----------
show
Whether to show the images.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
figs = {}
# plot the images
modules = {
"hu": self.ct_calibration_module,
"uniformity": self.uniformity_module,
"spatial resolution": self.spatial_resolution_module,
"low contrast": self.low_contrast_module,
}
for key, module in modules.items():
fig, ax = plt.subplots(**plt_kwargs)
module.plot(ax)
figs[key] = fig
# plot the one-off MTF image
fig, ax = plt.subplots(**plt_kwargs)
figs["mtf"] = fig
self.spatial_resolution_module.mtf.plot(ax)
# plot the side view
fig, ax = plt.subplots(**plt_kwargs)
figs["side"] = fig
self.plot_side_view(ax)
plt.tight_layout()
if show:
plt.show()
return figs
[docs]
def save_images(
self,
directory: Path | str | None = None,
to_stream: bool = False,
**plt_kwargs,
) -> list[Path | BytesIO]:
"""Save separate images to disk or stream.
Parameters
----------
directory
The directory to write the images to. If None, will use current working directory
to_stream
Whether to write to stream or disk. If True, will return streams. Directory is ignored in that scenario.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
figs = self.plot_images(show=False, **plt_kwargs)
paths = []
for name, fig in figs.items():
if to_stream:
path = io.BytesIO()
else:
destination = Path(directory) or Path.cwd()
path = (destination / name).with_suffix(".png").absolute()
fig.savefig(path)
paths.append(path)
return paths
[docs]
def find_phantom_roll(self, func=lambda roi: roi.bbox_area) -> float:
"""Determine the "roll" of the phantom.
Only difference of base method is that we sort the ROIs by size,
not by being in the center since the two we're looking for are both right-sided.
"""
return super().find_phantom_roll(func)
[docs]
def results(self) -> str:
"""Return the results of the analysis as a string. Use with print()."""
string = (
f"\n - ACR CT 464 QA Test - \n"
f"HU ROIs: {self.ct_calibration_module.roi_vals_as_str}\n"
f"Contrast to Noise Ratio: {self.low_contrast_module.cnr():2.2f}\n"
f"Uniformity ROIs: {self.uniformity_module.roi_vals_as_str}\n"
f'Uniformity Center ROI standard deviation: {self.uniformity_module.rois["Center"].std:2.2f}\n'
f"MTF 50% (lp/mm): {self.spatial_resolution_module.mtf.relative_resolution(50):2.2f}\n"
)
return string
def _generate_results_data(self) -> ACRCTResult:
return ACRCTResult(
phantom_model="ACR CT 464",
phantom_roll_deg=self.catphan_roll,
origin_slice=self.origin_slice,
num_images=self.num_images,
ct_module=CTModuleOutput(
offset=0,
roi_distance_from_center_mm=self.ct_calibration_module.roi_dist_mm,
roi_radius_mm=self.ct_calibration_module.roi_radius_mm,
roi_settings=self.ct_calibration_module.roi_settings,
rois={
name: roi.pixel_value
for name, roi in self.ct_calibration_module.rois.items()
},
),
uniformity_module=UniformityModuleOutput(
offset=CT_UNIFORMITY_MODULE_OFFSET_MM,
roi_distance_from_center_mm=self.uniformity_module.roi_dist_mm,
roi_radius_mm=self.uniformity_module.roi_radius_mm,
roi_settings=self.uniformity_module.roi_settings,
rois={
name: roi.pixel_value
for name, roi in self.uniformity_module.rois.items()
},
center_roi_stdev=self.uniformity_module.rois["Center"].std,
),
spatial_resolution_module=SpatialResolutionModuleOutput(
offset=CT_SPATIAL_RESOLUTION_MODULE_OFFSET_MM,
roi_distance_from_center_mm=self.spatial_resolution_module.roi_dist_mm,
roi_radius_mm=self.spatial_resolution_module.roi_radius_mm,
roi_settings=self.spatial_resolution_module.roi_settings,
rois={
name: roi.pixel_value
for name, roi in self.spatial_resolution_module.rois.items()
},
lpmm_to_rmtf=self.spatial_resolution_module.mtf.norm_mtfs,
),
low_contrast_module=LowContrastModuleOutput(
offset=CT_LOW_CONTRAST_MODULE_OFFSET_MM,
roi_distance_from_center_mm=self.low_contrast_module.roi_dist_mm,
roi_radius_mm=self.low_contrast_module.roi_radius_mm,
roi_settings=self.low_contrast_module.roi_settings,
rois={
name: roi.pixel_value
for name, roi in self.low_contrast_module.rois.items()
},
cnr=self.low_contrast_module.cnr(),
),
)
[docs]
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"{self._model} Analysis"
texts = [
" - ACR CT 464 Results - ",
f"HU Linearity ROIs: {self.ct_calibration_module.roi_vals_as_str}",
f"Low contrast visibility: {self.low_contrast_module.cnr():2.2f}",
f"Uniformity ROIs: {self.uniformity_module.roi_vals_as_str}",
]
analysis_images = self.save_images(to_stream=True)
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 idx, text in enumerate(texts):
canvas.add_text(text=text, location=(1.5, 23 - idx * 0.5))
for page, img in enumerate(analysis_images):
canvas.add_new_page()
canvas.add_image(img, location=(1, 5), dimensions=(18, 18))
canvas.finish()
if open_file:
webbrowser.open(filename)
def _module_offsets(self) -> list[float]:
absolute_origin_position = self.dicom_stack[self.origin_slice].z_position
relative_offsets_mm = [
0,
CT_UNIFORMITY_MODULE_OFFSET_MM,
CT_LOW_CONTRAST_MODULE_OFFSET_MM,
CT_SPATIAL_RESOLUTION_MODULE_OFFSET_MM,
]
return [
absolute_origin_position + offset_mm for offset_mm in relative_offsets_mm
]
class MRSlice11PositionModule(CatPhanModule):
common_name = "Slice Position, Slice 11"
roi_settings = {
"Left": {"width": 2, "height": 25, "distance": 65, "angle": 2.5},
"Right": {"width": 2, "height": 25, "distance": 65, "angle": -2.5},
}
rois: dict = {}
def _setup_rois(self) -> None:
for name, setting in self.roi_settings.items():
# angle is +90 because pointing right is 0, and these rois move downward, not rightward
self.rois[name] = RectangleROI(
self.image,
setting["width_pixels"],
setting["height_pixels"],
self.catphan_roll - 90 + setting["angle"],
setting["distance_pixels"],
self.phan_center,
)
@property
def bar_difference_mm(self) -> float:
"""The difference in height between the two angled bars"""
idxs = []
for roi in (self.rois["Right"], self.rois["Left"]):
prof = roi.pixel_array.max(axis=np.argmin(roi.pixel_array.shape))
mid_height = (prof.max() - prof.min()) / 2 + prof.min()
idx = find_nearest_idx(prof, mid_height)
idxs.append(idx)
return (idxs[0] - idxs[1]) * self.mm_per_pixel
@property
def slice_shift_mm(self) -> float:
"""The effective shift in phantom position in the S/I direction. Because bars are at 45 degrees, the shift is half the bar difference"""
return self.bar_difference_mm / 2
def plot_rois(self, axis: plt.Axes) -> None:
"""Plot the ROIs to the axis.
We overload because simple rectangle ROIs don't have a pass/fail color.
"""
for roi in self.rois.values():
roi.plot2axes(axis, edgecolor="blue")
[docs]
class MRSlice11ModuleOutput(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 #:
roi_settings: dict #:
rois: dict #:
bar_difference_mm: float #:
slice_shift_mm: float #:
class MRSlice1Module(CatPhanModule):
common_name = "Slice 1 (Thickness, Offset, Resolution)"
slice_lines: dict[str, Line]
thickness_rois: dict[str, ThicknessROI] = {}
thickness_roi_settings = {
"Top": {"width": 100, "height": 3, "distance": -3},
"Bottom": {"width": 100, "height": 3, "distance": 2.5},
}
roi_settings = {
"Row Reference": {"radius": 9, "distance": 58, "angle": 135, "lp/mm": 0},
"Col Reference": {"radius": 9, "distance": 58, "angle": 135, "lp/mm": 0},
"Row 1.1": {"radius": 3, "distance": 40, "angle": 116, "lp/mm": 1 / 1.1},
"Col 1.1": {"radius": 3, "distance": 44, "angle": 104, "lp/mm": 1 / 1.1},
"Row 1.0": {"radius": 3, "distance": 36, "angle": 81, "lp/mm": 1.0},
"Col 1.0": {"radius": 3, "distance": 44, "angle": 74, "lp/mm": 1.0},
"Row 0.9": {"radius": 2, "distance": 46, "angle": 52, "lp/mm": 1 / 0.9},
"Col 0.9": {"radius": 2, "distance": 55, "angle": 51, "lp/mm": 1 / 0.9},
}
position_roi_settings = {
"Left": {"width": 2, "height": 25, "distance": 65, "angle": 2.5},
"Right": {"width": 2, "height": 25, "distance": 65, "angle": -2.5},
}
position_rois: dict = {}
rois: dict[str, HighContrastDiskROI]
spacings = [0, 1 / 1.1, 1, 1 / 0.9]
def _setup_rois(self) -> None:
# thickness
for name, setting in self.thickness_roi_settings.items():
# angle is +90 because pointing right is 0, and these rois move downward, not rightward
self.thickness_rois[name] = ThicknessROI(
self.image,
setting["width_pixels"],
setting["height_pixels"],
self.catphan_roll + 90,
setting["distance_pixels"],
self.phan_center,
)
# spatial res
for name, setting in self.roi_settings.items():
self.rois[name] = HighContrastDiskROI(
self.image,
setting["angle_corrected"],
setting["radius_pixels"],
setting["distance_pixels"],
self.phan_center,
contrast_threshold=1.0, # fixed to 1 so everything passes. We aren't evaluating pass/fail here
)
# slice position
for name, setting in self.position_roi_settings.items():
# angle is +90 because pointing right is 0, and these rois move downward, not rightward
self.position_rois[name] = ThicknessROI(
self.image,
setting["width_pixels"],
setting["height_pixels"],
self.catphan_roll - 90 + setting["angle"],
setting["distance_pixels"],
self.phan_center,
)
def plot_rois(self, axis: plt.Axes) -> None:
for roi in self.position_rois.values():
roi.plot2axes(axis, edgecolor="blue")
for roi in self.thickness_rois.values():
roi.plot2axes(axis, edgecolor="blue")
for roi, mtf in zip(self.rois.values(), self.rois.values()):
roi.plot2axes(axis, edgecolor="g")
@property
def bar_difference_mm(self) -> float:
"""The difference in height between the two angled bars"""
left_array = self.position_rois["Left"].long_profile.values
left_mid_height = (left_array.max() - left_array.min()) / 2 + left_array.min()
left_idx = find_nearest_idx(left_array, left_mid_height)
right_array = self.position_rois["Right"].long_profile.values
right_mid_height = (
right_array.max() - right_array.min()
) / 2 + right_array.min()
right_idx = find_nearest_idx(right_array, right_mid_height)
return (right_idx - left_idx) * self.mm_per_pixel
@property
def slice_shift_mm(self) -> float:
"""The effective shift in phantom position in the S/I direction. Because bars are at 45 degrees, the shift is half the bar difference"""
return self.bar_difference_mm / 2
@property
def measured_slice_thickness_mm(self) -> float:
"""The slice thickness as determined by the two angled ROIs in the center of Slice 1"""
top = self.thickness_rois["Top"].wire_fwhm * self.mm_per_pixel
bottom = self.thickness_rois["Bottom"].wire_fwhm * self.mm_per_pixel
return 0.2 * (top * bottom) / (top + bottom)
@property
def row_mtf(self) -> MTF:
"""The MTF of the spatial resolution module looking at the row-wise ROIs"""
return MTF.from_high_contrast_diskset(
spacings=self.spacings,
diskset=list(roi for name, roi in self.rois.items() if "Row" in name),
)
@property
def col_mtf(self) -> MTF:
"""The MTF of the spatial resolution module looking at the column-wise ROIs"""
return MTF.from_high_contrast_diskset(
spacings=self.spacings,
diskset=list(roi for name, roi in self.rois.items() if "Col" in name),
)
[docs]
class MRSlice1ModuleOutput(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 #:
roi_settings: dict #:
rois: dict #:
bar_difference_mm: float #:
slice_shift_mm: float #:
measured_slice_thickness_mm: float #:
row_mtf_50: float #:
col_mtf_50: float #:
class MRUniformityModule(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.
"""
common_name = "Signal Uniformity"
roi_settings = {
"Center": {
"angle": 90,
"distance": 5,
"radius": 80,
}, # 80 radius ~= 200cm2, per the manual
}
ghost_roi_settings = {
# size of ~900mm2 per the manual
"Top": {"angle": -90, "distance": 110, "width": 60, "height": 15},
"Bottom": {"angle": 90, "distance": 110, "width": 60, "height": 15},
"Left": {"angle": 180, "distance": 110, "width": 15, "height": 60},
"Right": {"angle": 0, "distance": 110, "width": 15, "height": 60},
}
ghost_rois: dict = {}
def __init__(self, catphan, offset):
self.tesla = float(catphan.dicom_stack.metadata.MagneticFieldStrength)
super().__init__(catphan, tolerance=None, offset=offset)
def _setup_rois(self) -> None:
super()._setup_rois()
for name, roi in self.ghost_roi_settings.items():
self.ghost_rois[name] = RectangleROI(
self.image,
roi["width_pixels"],
roi["height_pixels"],
roi["angle"] + self.catphan_roll,
roi["distance_pixels"],
self.phan_center,
)
def plot_rois(self, axis: plt.Axes) -> None:
super().plot_rois(axis)
for roi in self.ghost_rois.values():
roi.plot2axes(axis, edgecolor="yellow")
@property
def percent_image_uniformity(self) -> float:
"""PIU value calculated via section 5.3 of the manual"""
piu_high = np.percentile(self.rois["Center"].pixel_values, 99)
piu_low = np.percentile(self.rois["Center"].pixel_values, 1)
return 100 * (1 - ((piu_high - piu_low) / (piu_high + piu_low)))
@property
def piu_passed(self) -> bool:
"""Section 5.4"""
if self.tesla < 3:
return self.percent_image_uniformity > 85
else:
return self.percent_image_uniformity > 80
@property
def ghosting_ratio(self) -> float:
"""Ghosting ratio of section 6.3 of the manual"""
top = self.ghost_rois["Top"].pixel_value
bottom = self.ghost_rois["Bottom"].pixel_value
left = self.ghost_rois["Left"].pixel_value
right = self.ghost_rois["Right"].pixel_value
return abs(
((top + bottom) - (left + right)) / (2 * self.rois["Center"].pixel_value)
)
@property
def psg(self) -> float:
"""Percent Signal Ghosting"""
return self.ghosting_ratio * 100
@property
def psg_passed(self) -> bool:
"""Whether the PSG is within tolerance"""
return self.psg < 3.0
class GeometricDistortionModule(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.
"""
common_name = "Geometric Distortion"
profiles: dict
def _setup_rois(self) -> None:
self.profiles = {}
bin_image = self.image.as_binary(threshold=np.percentile(self.image, 60))
bin_image = ndimage.binary_fill_holes(bin_image).astype(float)
# calculate horizontal
data = bin_image[int(self.phan_center.y), :]
prof = FWXMProfilePhysical(values=data, dpmm=1 / self.mm_per_pixel)
line = Line(
Point(prof.field_edge_idx(side="left"), self.phan_center.y),
Point(prof.field_edge_idx(side="right"), self.phan_center.y),
)
self.profiles["horizontal"] = {
"width (mm)": prof.field_width_mm,
"line": line,
}
# calculate vertical
data = bin_image[:, int(self.phan_center.x)]
prof = FWXMProfilePhysical(values=data, dpmm=1 / self.mm_per_pixel)
line = Line(
Point(self.phan_center.x, prof.field_edge_idx(side="left")),
Point(self.phan_center.x, prof.field_edge_idx(side="right")),
)
self.profiles["vertical"] = {
"width (mm)": prof.field_width_mm,
"line": line,
}
# calculate negative diagonal
# calculate slope equation intercept
# b = y - (+1)x
b = self.phan_center.y - self.phan_center.x
xs = np.arange(0, self.image.shape[1])
ys = xs + b
coords = ndimage.map_coordinates(bin_image, [ys, xs], order=1, mode="mirror")
# pixels are now diagonal and thus spacing between pixels is now the hypotenuse
prof = FWXMProfilePhysical(
values=coords, dpmm=1 / (self.mm_per_pixel * math.sqrt(2))
)
line = Line(
Point(
xs[int(round(prof.field_edge_idx(side="left")))],
ys[int(round(prof.field_edge_idx(side="left")))],
),
Point(
xs[int(round(prof.field_edge_idx(side="right")))],
ys[int(round(prof.field_edge_idx(side="right")))],
),
)
self.profiles["negative diagonal"] = {
"width (mm)": prof.field_width_mm,
"line": line,
}
# calculate positive diagonal
# calculate slope equation intercept
# b = y - (-1)x
b = self.phan_center.y + self.phan_center.x
ys = -xs + b
coords = ndimage.map_coordinates(bin_image, [ys, xs], order=1, mode="mirror")
prof = FWXMProfilePhysical(
values=coords, dpmm=1 / (self.mm_per_pixel * math.sqrt(2))
)
line = Line(
Point(
xs[int(round(prof.field_edge_idx(side="left")))],
ys[int(round(prof.field_edge_idx(side="left")))],
),
Point(
xs[int(round(prof.field_edge_idx(side="right")))],
ys[int(round(prof.field_edge_idx(side="right")))],
),
)
self.profiles["positive diagonal"] = {
"width (mm)": prof.field_width_mm,
"line": line,
}
def plot_rois(self, axis: plt.Axes):
for name, profile_data in self.profiles.items():
profile_data["line"].plot2axes(axis, width=2, color="blue")
def distances(self) -> dict:
"""The measurements of the phantom size for all 4 lines in mm"""
return {name: f"{p['width (mm)']:2.2f}mm" for name, p in self.profiles.items()}
[docs]
class MRGeometricDistortionModuleOutput(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 #:
profiles: dict #:
distances: dict #:
[docs]
class ACRMRIResult(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."""
phantom_model: str #:
phantom_roll_deg: float #:
origin_slice: int #:
num_images: int #:
slice1: MRSlice1ModuleOutput #:
slice11: MRSlice11ModuleOutput #:
uniformity_module: MRUniformityModuleOutput #:
geometric_distortion_module: MRGeometricDistortionModuleOutput #:
[docs]
class ACRMRILarge(CatPhanBase, ResultsDataMixin[ACRMRIResult]):
_model = "ACR MRI Large"
catphan_radius_mm = 100
min_num_images = 4
air_bubble_radius_mm = 20
slice1 = MRSlice1Module
geometric_distortion = GeometricDistortionModule
uniformity_module = MRUniformityModule
slice11 = MRSlice11PositionModule
[docs]
def plot_analyzed_subimage(self, *args, **kwargs):
raise NotImplementedError("Use `plot_images`")
[docs]
def save_analyzed_subimage(self, *args, **kwargs):
raise NotImplementedError("Use `save_images`")
[docs]
def localize(self) -> None:
self._phantom_center_func = self.find_phantom_axis()
self.catphan_roll = self.find_phantom_roll()
# 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 cover the extent of module configuration. "
"This means not all modules were included in the scan. Rescan the phantom to include all "
"relevant modules, or change the offset values."
)
def _module_offsets(self) -> list[float]:
absolute_origin_position = self.dicom_stack[self.origin_slice].z_position
relative_offsets_mm = [
0,
MR_GEOMETRIC_DISTORTION_MODULE_OFFSET_MM,
MR_UNIFORMITY_MODULE_OFFSET_MM,
MR_SLICE11_MODULE_OFFSET_MM,
]
return [
absolute_origin_position + offset_mm for offset_mm in relative_offsets_mm
]
[docs]
def find_phantom_roll(self) -> float:
"""Determine the "roll" of the phantom. This algorithm uses the circular left-upper hole on slice 1 as the reference
Returns
-------
float : the angle of the phantom in **degrees**.
"""
# get edges and make ROIs from it
slice = Slice(self, self.origin_slice)
larr, regions, _ = get_regions(slice)
try:
# find appropriate ROIs and grab the two most centrally positioned ones
circle_bubbles = [
r
for r in regions
if (self._is_right_area(r) and self._is_right_eccentricity(r))
]
exact_size = np.pi * ((self.air_bubble_radius_mm / self.mm_per_pixel) ** 2)
most_similar_bubble = sorted(
circle_bubbles, key=lambda r: abs(r.filled_area - exact_size)
)[0]
y_dist = most_similar_bubble.centroid[0] - slice.phan_center.y
x_dist = most_similar_bubble.centroid[1] - slice.phan_center.x
phan_roll = np.arctan2(y_dist, x_dist)
corrected_roll = (
np.rad2deg(phan_roll) + 135
) # bubble is at top-left. perfect placement is -135
return corrected_roll
except Exception:
raise RuntimeError(
"Could not determine the roll of the phantom. Ensure the 20mm top-left circle is visible on Slice 1"
)
[docs]
def analyze(self, echo_number: int | None = None) -> None:
"""Analyze the ACR CT phantom
Parameters
----------
echo_number:
The echo to analyze. If not passed, uses the minimum echo number found.
"""
self._select_echo_images(echo_number)
self.localize()
self.slice1 = self.slice1(self, offset=0)
self.geometric_distortion = self.geometric_distortion(
self, offset=MR_GEOMETRIC_DISTORTION_MODULE_OFFSET_MM
)
self.uniformity_module = self.uniformity_module(
self, offset=MR_UNIFORMITY_MODULE_OFFSET_MM
)
self.slice11 = self.slice11(self, offset=MR_SLICE11_MODULE_OFFSET_MM)
def _select_echo_images(self, echo_number: int | None) -> None:
"""Select out the images that match the given echo number"""
# we check for multiple echos. We only pick the first echo found.
# this is probably not the best logic but we somehow have to pick
# Echo Numbers is an int; https://dicom.innolitics.com/ciods/mr-image/mr-image/00180086
# in case EchoNumbers isn't there, use all
try:
all_echos = {int(i.metadata.EchoNumbers) for i in self.dicom_stack}
except AttributeError:
# no manipulation; use all images
return
if echo_number is None:
echo_number = min(all_echos)
if len(all_echos) > 1:
warnings.warn(
f"Multiple echoes found ({all_echos}) and no echo number was passed. Using echo # {echo_number}"
)
if echo_number not in all_echos:
raise ValueError(
f"Echo number {echo_number} was passed but not found in the dataset. Found echo numbers: {all_echos}. Remove the echo_number parameter or pick a valid echo number."
)
# drop images that don't have the same echo number
to_pop = []
for idx, img in enumerate([i for i in self.dicom_stack].copy()):
if int(img.metadata.EchoNumbers) != echo_number:
to_pop.append(idx)
for idx in sorted(to_pop, reverse=True):
del self.dicom_stack[idx]
del self.dicom_stack.metadatas[idx]
[docs]
def plot_analyzed_image(self, show: bool = True, **plt_kwargs) -> plt.Figure:
"""Plot the analyzed image
Parameters
----------
show
Whether to show the image.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
# set up grid and axes
fig = plt.figure(**plt_kwargs)
grid_size = (2, 3)
slice1_ax = plt.subplot2grid(grid_size, (0, 0))
self.slice1.plot(slice1_ax)
geom_ax = plt.subplot2grid(grid_size, (0, 1))
self.geometric_distortion.plot(geom_ax)
unif_ax = plt.subplot2grid(grid_size, (0, 2))
self.uniformity_module.plot(unif_ax)
position_ax = plt.subplot2grid(grid_size, (1, 0))
self.slice11.plot(position_ax)
side_view_ax = plt.subplot2grid(grid_size, (1, 1))
self.plot_side_view(side_view_ax)
spatial_res_graph = plt.subplot2grid(grid_size, (1, 2))
self.slice1.row_mtf.plot(spatial_res_graph, label="Row-wise rMTF")
self.slice1.col_mtf.plot(spatial_res_graph, label="Column-wise rMTF")
spatial_res_graph.legend()
# finish up
plt.tight_layout()
if show:
plt.show()
return fig
[docs]
def plot_images(self, show: bool = True, **plt_kwargs) -> dict[str, plt.Figure]:
"""Plot all the individual images separately
Parameters
----------
show
Whether to show the images.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
figs = {}
# plot the images
modules = {
"geometric": self.geometric_distortion,
"slice 1": self.slice1,
"signal uniformity": self.uniformity_module,
"slice 11": self.slice11,
}
for key, module in modules.items():
fig, ax = plt.subplots(**plt_kwargs)
module.plot(ax)
figs[key] = fig
# plot rMTF
fig, ax = plt.subplots(**plt_kwargs)
self.slice1.row_mtf.plot(ax, label="Row-wise rMTF")
self.slice1.col_mtf.plot(ax, label="Column-wise rMTF")
ax.legend()
figs["rMTF"] = fig
# plot the side view
fig, ax = plt.subplots(**plt_kwargs)
figs["side"] = fig
self.plot_side_view(ax)
if show:
plt.show()
return figs
def _detected_modules(self) -> list[CatPhanModule]:
return [
self.slice1,
self.slice11,
self.uniformity_module,
self.geometric_distortion,
]
[docs]
def save_images(
self,
directory: Path | str | None = None,
to_stream: bool = False,
**plt_kwargs,
) -> list[Path | BytesIO]:
"""Save separate images to disk or stream.
Parameters
----------
directory
The directory to write the images to. If None, will use current working directory
to_stream
Whether to write to stream or disk. If True, will return streams. Directory is ignored in that scenario.
plt_kwargs
Keywords to pass to matplotlib for figure customization.
"""
figs = self.plot_images(show=False, **plt_kwargs)
paths = []
for name, fig in figs.items():
if to_stream:
path = io.BytesIO()
else:
destination = Path(directory) or Path.cwd()
path = (destination / name).with_suffix(".png").absolute()
fig.savefig(path)
paths.append(path)
return paths
[docs]
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"{self._model} Analysis"
analysis_images = self.save_images(to_stream=True)
canvas = pdf.PylinacCanvas(
filename, page_title=analysis_title, metadata=metadata
)
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))
shortened_texts = [
textwrap.wrap(r, width=110) for r in self.results(as_str=False)
]
idx = 0
for items in enumerate(shortened_texts):
for text in items:
canvas.add_text(text=text, location=(1.5, 25 - idx * 0.5))
idx += 1
for page, img in enumerate(analysis_images):
canvas.add_new_page()
canvas.add_image(img, location=(1, 5), dimensions=(18, 18))
canvas.finish()
if open_file:
webbrowser.open(filename)
[docs]
def results(self, as_str: bool = True) -> str | tuple:
"""Return the results of the analysis as a string. Use with print()."""
string = (
f" - {self._model} Results - ",
f"Geometric Distortions: {self.geometric_distortion.distances()}",
f"Slice Thickness: {self.slice1.measured_slice_thickness_mm:2.2f}mm",
f"Slice 1 S/I Position shift: {self.slice1.slice_shift_mm:2.2f}mm",
f"Slice 11 S/I Position shift: {self.slice11.slice_shift_mm:2.2f}mm",
f"Uniformity PIU: {self.uniformity_module.percent_image_uniformity:2.2f}",
f"Percent-signal ghosting: {self.uniformity_module.psg:2.2f}%",
f'Uniformity Center ROI standard deviation: {self.uniformity_module.rois["Center"].std:2.2f}',
f"Row-wise MTF 50% (lp/mm): {self.slice1.row_mtf.relative_resolution(50):2.2f}",
f"Column-wise MTF 50% (lp/mm): {self.slice1.col_mtf.relative_resolution(50):2.2f}",
)
if as_str:
return "\n".join(string)
else:
return string
def _generate_results_data(self) -> ACRMRIResult:
return ACRMRIResult(
phantom_model=self._model,
phantom_roll_deg=self.catphan_roll,
origin_slice=self.origin_slice,
num_images=self.num_images,
slice1=MRSlice1ModuleOutput(
offset=0,
roi_settings=self.slice1.roi_settings,
rois=rois_to_results(self.slice1.rois),
bar_difference_mm=self.slice1.bar_difference_mm,
slice_shift_mm=self.slice1.slice_shift_mm,
measured_slice_thickness_mm=self.slice1.measured_slice_thickness_mm,
row_mtf_50=self.slice1.row_mtf.relative_resolution(50),
col_mtf_50=self.slice1.col_mtf.relative_resolution(50),
),
slice11=MRSlice11ModuleOutput(
offset=MR_SLICE11_MODULE_OFFSET_MM,
bar_difference_mm=self.slice11.bar_difference_mm,
slice_shift_mm=self.slice11.slice_shift_mm,
rois=rois_to_results(self.slice11.rois),
roi_settings=self.slice11.roi_settings,
),
geometric_distortion_module=MRGeometricDistortionModuleOutput(
offset=MR_GEOMETRIC_DISTORTION_MODULE_OFFSET_MM,
profiles=self.geometric_distortion.profiles,
distances=self.geometric_distortion.distances(),
),
uniformity_module=MRUniformityModuleOutput(
offset=0,
roi_settings=self.uniformity_module.roi_settings,
rois=rois_to_results(self.uniformity_module.rois),
ghost_roi_settings=self.uniformity_module.ghost_roi_settings,
ghost_rois=rois_to_results(self.uniformity_module.ghost_rois),
psg=self.uniformity_module.psg,
ghosting_ratio=self.uniformity_module.ghosting_ratio,
piu=self.uniformity_module.percent_image_uniformity,
piu_passed=self.uniformity_module.piu_passed,
),
)