qs package has been removed from dependencies (Suggests).
Users with legacy .Rdcm files (pre-version 1.11.1) will now receive a detailed
message guiding them to install qs from GitHub or CRAN Archives if needed.
Migration to the new qs2 format via Rdcm.upgrade() is strongly recommended
while using R <= 4.4.The qs package, used to create Rdcm files, will soon be deprecated. The espadon package now uses qs2, but allows loading of older *.Rdcm files, with load.T.MAT(), load.obj.from.Rdcm(), load.patient.from.Rdcm(), load.Rdcm.raw.data(). Rdcm files can be updated with the Rdcm_upgrade() function.
DICOM MR export is now possible with the export() function.
struct.update_roiinfo() updates the roi.info element in the espadon object of the struct class when point coordinates have been modified in the roi.data element.
err.metrics.from.bin() is added to complement the err.metrics.from.roi() function. It is used to calculate errors on binary selections.
Execution speed of dicom.raw.data.anonymizer() and dicom.patient.anonymiser() has been increased.
Min and max values in err.metrics.from.roi() have been added.
The dicom.viewer() cursor update has been modified.
detection of Rdcm version in the load_Rcm_raw_data() function and all functions that use it, such as load.obj.from.Rdcm, load.patient.from.Rdcm
The error in the load.obj.from.Rdcm() function when loading a single contour among others has been corrected.
err.metrics.from.roi() calculates various metrics (ME, MAE, MSE, RMSE) to compare 2 “volume” class objects in the zones delimited by requested RoI.
vol.error() and vol.abserror() provide the error and the absolute error volume between 2 volumes.
nesting.bin() and nesting.cube() corrected for 2D use.
plot.struct() and display.plane() have been corrected for RoI selection, when some RoIs do not contain data.
get.value.from.ijk() and all functions that depend on it (vol.regrid(), bin.from.roi() for example) has been corrected in the counting of subvoxels contained in a new voxel (and condition for reaching these voxels) in 'average' mode.
calculation of std value in rt.indices.from.roi() and rt.indices.from.nin() is corrected.
Voxel values calculated by vol.regrid() in “Average” mode are corrected.
Correction of file name creation error in the export() function.
bin.cuboid(), bin.cylinder() and bin.ellipsoid() create espadon objects of class “volume”, and of modality “binary” or “weight”, by selecting the voxels defining a rectangular cuboid, an elliptical cylinder or an ellipsoid.
add.shape() adds the shape defined by espadon volume object of the modality "binary" or "weight" to a 3D volume.
The execution time of the bin.opening(), bin.closing(), bin.dilation() and bin.erosion() functions has been improved.
vol.create() now accepts random Gaussian values.
export() now manages espadon object fields with the value NA. RT-struct export writes the IUD from each CT slice to each ROI contour slice.
nesting.bin() now takes the xyz.margin argument into account, for mesh class objects.
bin.from.roi() now handles the “weight” modality, in which the value of each voxel is its volume fraction included in the ROI. In the “binary” modality, the voxel whose fraction is greater than 0.5 is set to TRUE. The emergence of this new modality impacts histo.from.roi(), histo.from.bin(), rt.indices.from.roi() and rt.indices.from.bin() functions. The use of the “weight” modality enables us to get closer to the ROI contours. This slightly modifies the results compared with the previous version.
histo.from.roi() and histo.from.bin() return a dataframe of dose-volume histograms (in percent) for ROIs if requested.
nesting.bin() now works on meshes.
bin.closing(), bin.erosion(), bin.dilation(), bin.opening (), nesting.roi(), add.margin(),get.value.from.ijk now stop instead of warning and return NULL
enhancement of struct.clustering example.
display.plane() now stops directly on error. It takes into account the argument 'struct.dxyz'.
vol.subsampling(): The xyz-ranges error has been corrected.
plot.mesh() can now display opened meshes
plot.volume(), in case of add=FALSE, has its map range rectified.
load.obj.from.dicom() and load.patient.from.dicom() no longer uses the tag ( 0020,0012) to separate DICOM objects
dicom.patient.anonymiser() anonymises all DICOM files in a patient's directory.
export() function exports struct class objects and volume class objects with CT or RTDOSE modality in DICOM format.
bin.from.roi() : back to the more robust version 1.5.1.
load.obj.from.dicom(), load.obj.from.Rdcm() and load.obj.data() modify their strategy for calculating the thickness between polygons in DICOM rt-struct files: all z-slices in each ROI are taken into account to calculate all thicknesses. The resulting thickness will be the median of all these thicknesses.
dicom_set_tag_value(): correction of value acceptance for tags with "CS" VR.
plot() function to plot in 2D espadon objects of class 'volume', 'struct' and 'mesh'.
pal.rainbow() to produce a color palette suitable for displaying the dose in overlay.
display.plane() runs faster in sagital and frontal view, and accepts ... arguments. If the color palette is the RVV palette, breaks are automatically calculated between -1000 and 1000, as with the display.kplane function.
bin.from.roi() runs faster. This affects the functions histo.from.roi(), struct.clustering() and rt.indices.from.roi().
sp.similarity.from.mesh() now computes surface Added Path Length.
load.patient.from.dicom(), load.obj.from.dicom() and dicom.to.Rdcm.converter() now correctly handle the 'scan spot meterset weight' of rt ion plan modality.
sp.similarity.from.mesh() now computes correctly sDSC metrics.
struct.from.mesh() now correctly calculates complex mesh contours, which depend on curvature.
sp.similarity.from.bin() to calculate spatial similarity such as volumetric Dice similarity or Dice-Jaccard coefficient, from binary volumes
sp.similarity.from.mesh() to calculate spatial similarity such as Hausdorff distances and surface Dice similarity coefficient, from meshes.
display.kplane() and display.plane() now display images correctly with MAC OS.
nesting.cube() : Rectification of points included in the cube and located at its edges.
load.patient.from.dicom(), load.obj.from.dicom() and dicom.to.Rdcm.converter() now handle special characters as soon as DICOM files are decrypted. The functions load.Rdcm.raw.data() and Rdcm.upgrade() functions have been updated accordingly.
load.patient.from.dicom() and dicom.to.Rdcm.converter() now take into account all rtdose even those without tag (300C,0006). The import of tag content (3006,00A6) from rtstruct files has been corrected.
dicom_set_tag_value() now accepts "IS" DICOM value representation with many integers.
struct.from.bin() and struct.from.roi() create a espadon struct object with roi.obs$label equal to "".
get.extreme.pt() now takes the max argument into account.
obj.create() to create an espadon object with the essential properties it must have.
struct.from.mesh() to create a struct object with a unique RoI, defined by the contours of a mesh.
display.plane() displays now point or open.planar contour, provided that all points of the contour are located in the same plane.
mesh.from.bin() now has default tol argument equal to min(bin$dxyz)/2, to limit meshing errors.
ref.srctodest.add() only warns and does not stop when the link between src.ref and dest.ref is already defined in the T.MAT object.
vol.regrid() : The voxels of the volume edges are calculated.
vol.oversampling() now makes the center of the volume invariant.
get.extreme.pt() : management of NA value if obj is an rtstruct.
struct.merge() : roi.obs field is reducted to selected RoI. The RoI number is incremented in roi.obs and roi.info fields
gamma.index() : the 1 voxel shift has been corrected
rtplan loading : correction of the calculation of the returned value $beam.orientation for functions that can load an object of class rtplan, like load.data.from.patient(),load.patient.from.Rdcm(), load.obj.from.dicom(), load.obj.from.Rdcm() and load.obj.data().
struct loading : All functions that can load an object of class struct, like load.data.from.patient(),load.patient.from.Rdcm(), load.obj.from.dicom(), load.obj.from.Rdcm() and load.obj.data() now calculate better the thickness between planes of the contours when the link with the reference imaging object has not been done. When this link is made, the thickness is now equal to the interplan of the reference object. This change impacts the calculation of volumes from contours, of DVH, and rt.indices.from.roi.
vol.oversampling() and vol.subsampling() handles now the modality 'binary'
display.obj.links() : correction of the color management of the different frames of reference.
study.deployment() to deploy DICOM files from multiple patients. This function simplifies the analysis of multi-center or multi-expert studies in dosimetry challenges, contouring consensus searches, etc.
vol.repair() to repair missing planes in volumes.
get.value.from.mesh() to retrieve the values of an object of class "volume" at the desired depth of a surface described by a mesh.
Rdcm.upgrade() to update Rdcm files that were created with a previous version.
set.reference.obj() to add to an espadon object the information identifying the espadon object from which it derives
The patient.name field, in addition to the PIN, is now handled by espadon objects. It impacts all functions that create or load espadon objects, like load.patient.from.dicom, load.patient.from.Rdcm(), vol.create(), dicom.to.Rdcm.converter(), load_Rdcm_raw_data() ...
For all functions with the alias argument, when specified, the returned object will include the reference fields ref.object.alias and ref.object.info, which specify from which objects it was derived.
load.patient.from.Rdcm() and load.TMAT() now accept a vector of Rdcm files as argument.
New argument ignore.duplicates appears in dicom.to.Rdcm.converter(), load.patient.from.dicom() and load.patient.from.Rdcm(). It allows to ignore identical DICOM objects (with the same SOP UID), but which have been registered with different protocols.
struct.in.new.ref() no longer recalculates the geometric information on the RoI, in case the new reference frame is identical to the old one.
ref.cut.plane.add() now has new default value for the origin argument.
get.extreme.pt() now returns the coordinates of the extreme points of the espadon objects of the mesh class, the struct class, the volume class, or from a binary selection .
nesting.cube() function restricts the espadon objects of the mesh class, in addition to the espadon objects of the volume class, to the cube defined by its two extreme vertices.
rt.gamma.index() and rt.chi.index() now correct the pixel size to match that found if the sensor was at the isocenter, in the case where the objects to be compared are of rtimage modality.
get.value.from.xyz() now sends an error if the xyz argument is not a vector of length 3 or a 3-column matrix or dataframe.
histo.from.roi() now has the over.sampling.factor argument to oversample the volume voxels before treatment for a better result.
the 'dcm' argument of dicom.parser() and dicom.viewer() can now be either a dicom file name, or a Rdcm file name or dicom raw data.
orientation.create() now provides orientation composed of unitary vectors.
display.ob.links() gives error-free information even if the exclusion argument is non-NULL.
save.to.Rdcm() now saves the espadon version.
load.patient.from.dicom() now issues a warning when espadon cannot import one of several DICOM objects, instead of crashing.
get.rigid.M() now returns diag(4) when source reference frame is equal to
destination reference frame, regardless T.MAT. It therefore has an impact on
all functions that use a T.MAT to change the reference frame.
rt.indices.from.roi no longer uses all RoI containing "ptv" as target RoI by default.
rt.gamma.index() and rt.chi.index() now details, in the gamma.info field or the chi.info field, the number of dose points, the number of evaluated dose points, the rate of evaluated dose points.
load.patient.from.dicom() and load.patient.from.Rdcm() no longer fail when the dicom.dvh file has no region of interest as reference.
get.obj.connection() now provides an error-free components matrix.
display.obj.links() now have a better representation of dicom.dvh
rt.incices.from.roi, and hence rt.indices.from.bin no longer crash when no volume information is requested.
vol.create() now has a well calculated midpoint.
New struct.clustering() function to create a new volume in which voxels are clustered and labeled by Regions of Interest (RoI) defined in an rt-struct.
New get.roi.connection() function to describe the interconnections between RoI, from an imaging volume of "cluster" modality.
New fan.beam(), fan.planar(), fan.sphere() functions to create different ray patterns: fan of rays uniformly distributed by angle in a pyramid, or fan of rays passing through all pixels of a plane, or fan of rays regularly or randomly equidistributed in a sphere. A new espadon class object 'fan' is created, described in espadon.class() help.
New fan.to.voxel() function to computes the indices of voxels crossed by a fan.
New nesting.bin() function to restrict a volume to the rectangular parallelepiped circumscribed to the selected voxels.
New vol.median() function to apply a 3D median filter on a volume.
New functions rt.gamma.index() and rt.chi.index() to compute respectively the local or global Gamma and Chi 2D/3D indices from a measurement and a reference.
New get.obj.connection() function to describe the connections between the DICOM objects of the patient.
load.patient.from.dicom(), and hence dicom.to.Rdcm.converter() and load.object.from.dicom(), increase the information coming from a DICOM rt-plan:, in particular, those of the beam control points, the position of the source and the direction of the beam for each control point. The new information are listed in espadon.class() function. They now also generate more useful warnings to inform the user of problems with their data, and avoid uninformative warnings.
display_obj_links() now differentiates the origin (by arrows) and the links (by lines).
load.Rdcm.raw.data() now detects espadon version changes.
get.value.from.ijk() now calculates the value when the volume contains only one plane.
vol.regrid() now allows resampling when volume planes are missing.
display.3D.contour() to allow the display of RoI (Region of Interest) of type point.
display.plane() now allows the display of volumes with only one plane, the display of RoI of type point and the display of overlays with different reference frames.
toy.load.patient() now generates data faster when an rt-dose is requested.
rt.indices.from.roi() and rt.indices.from.bin() now compute dosimetry index D.xcc
display.plane() and display.kplane() now make the values outside the color palette transparent when sat.transp = TRUE and not when FALSE.
load.T.MAT(), and hence load.patient.from.dicom() and dicom.to.Rdcm.converter(), no longer make errors even if the patient's source DICOM files do not give the same information about their PIN, birthday and sex. A simple warning about the uniqueness of the patient is generated.
vol.gradient() now returns volume with border values at NA.
grid.equal(), which compares the grids of two volumes, now returns TRUE even if grids are almost equal (difference due to rounding error).