Tag | (0022,0032) |
---|---|
Type | Required (1) |
Keyword | ReferenceCoordinates |
Value Multiplicity | 2-2n |
Value Representation | Single (FL) |
Image coordinates for the points on the referenced image that correspond to the points on this frame. See Section C.8.17.10.1.1.
Multiple pairs of values where the first value of each pair is the row and the second value of each pair is the column. Row is the vertical offset and column is the horizontal offset.
Specified with sub-pixel resolution such that the origin at the Top Left Hand Corner (TLHC) of the TLHC pixel is 0.0\0.0, the Bottom Right Hand Corner (BRHC) of the TLHC pixel is 1.0\1.0, and the BRHC of the BRHC pixel is Rows\Columns (see Figure C.10.5-1, except that row and column order is reversed). The values must be within the range 0\0 to Rows\Columns.
Part of the Ophthalmic Frame Location Functional Group Macro with usage: U
A frame can be described in terms of its "position" on another image. In the case of ophthalmic tomography (OPT) images with longitudinal orientation, the OPT image corresponds to a vertical slice along a trajectory on the referenced image. Each column of the OPT frame is nominally perpendicular to the image. Therefore, each column of the OPT frame can also correspond to a particular pixel on an associated image. It is normal in OPT analysis to indicate the location of the OPT slice by showing the corresponding pixel locations on the image. For Transverse frames, the position is indicated by two corners of a rectangle.
The simple geometric relationship Modules are not practical to use for describing the relationship between OPT and retinal photography images. The retinal surface is highly curved, and the optical path is influenced by the lens and other eye structures. It is impractical to measure all of the geometric and optical properties of the eye to the degree needed to establish the relationship by geometric computation. It is easy to capture a retinal image and the slice path on that image as part of the OPT acquisition.
The OPT columns for a retinal slice are not strictly perpendicular to the retina and there are minor systematic distortions as a result. This DICOM Functional Group does not attempt to capture the data needed to measure or correct for these effects, except for the presence of the Maximum Along-scan Distortion Attribute (0022,0038).
The relationship between the columns in the image frame and the reference image are illustrated in Figure C.8.17.10-1.
Frames can be captured using a linear scan with uniform column spacing. This permits the column position to be described by giving the endpoints of the line segment that was scanned. Figure C.8.17.10-1 shows such a scan labeled as "OCT Frame 1". When describing this frame using the Linear Column Locations Functional Group Macro, the Sequence Attributes would contain:
The Attributes of the Image SOP Instance Reference Macro would point to a frame of the reference image.
Reference Coordinates (0022,0032) would contain values that give the row and column coordinates for the image pixel on the referenced image that corresponds to the first column of the OPT Frame 1 and the row and column coordinates that corresponds to the last column of the OPT Frame 1.
Frames can also be taken along curved paths, or with non-uniform spacing. These frames cannot be described by just the two endpoints. These frames are described by specifying the referenced image pixel coordinates for each column in the frame. The frame labeled "OCT Frame 2" illustrates this kind of relationship. The Non-linear Column Locations Functional Group Macro description for this frame would contain:
The Attributes of the Image SOP Instance Reference Macro would point to a frame of the reference image.
Reference Coordinates (0022,0032) would contain 2N integer values. Their contents would be:
1-2 (e,f)
… other values
2L-1, 2L (g,h)
… other values
2N-1, 2N (x,y)
A three dimensional OPT raster scan image would be described by a series of Linear Column Location Sequences, one for each frame.
One expected use of column location is the automatic generation of annotation graphics display. The slice location can be indicated and dynamically controlled on the reference image while examining the observed slices. This replaces the relatively inflexible practice of creating a modified retinal image with burned in annotation to indicate the slice locations.
Figure C.8.17.10-1. Relationship of Longitudinal (Nominally Perpendicular) OCT Frames to Reference Frame
In the case of ophthalmic tomography (OPT) images with a transverse orientation, the OPT image corresponds to a sub-rectangle of the reference image. See Figure C.8.17.10-2. The OPT frame is nominally parallel to the reference image. Therefore, each corner of the OPT frame can also correspond to a particular pixel on an associated image. It is assumed that the edges of OPT frame are exactly parallel to the edges of the reference image.
Figure C.8.17.10-2. Relationship of Transverse (Nominally Parallel) OCT Frame to Reference Frame
Figure C.10.5-1. Sub-pixel Addressing Units in PIXEL Space