Programs for Needle Geometries

Rob MacLeod


1 Introduction

The programs described in this document all serve one purpose--to make management of needle geometry as easy as possible. They take care of all the steps required to convert digitized rod entry points and rod tips into geometry files, with a few additional ``goodies'' for fun.

The steps required to manage needles are as follows:

  1. Either a) remove the needles from the heart and replace each with a rod, making sure to keep track of the needle number, or b) leave needles in place.
  2. If a), digitize the locations of all rod tips and the rod entry points; if b) digitize the locations of the ends of the needles, at the point where the wires connect, and at the entry point of the needle.
  3. Convert the resulting file to a landmarks file with digtoxyz (See Section 2). Optionally, view the landmarks file in the context of the surface geometry, if it is available, with map3d.
  4. Convert the resulting landmarks file to a geometry file using needlegeom (see Section 3) and view the results again with map3d.
  5. Digitize the innermost surface of the geometry using map3d.
  6. Propagate this triangularization to all other surfaces in the geometry using needletri (see Section 4)
  7. View the result with map3d
  8. If you wish to have subsets of needles from the full geometry, use the program makeplanes (see Section 5).

The programs are not perfect, nor is the documentation, but we can make it that way with your help. Let Rob know what is not clear or what fails to perform as required and we will aim to please!


2 Processing digitizer files: digtoxyz

The format of data that comes from the digitizer is not directly usable by any of the standard CVRTI programs and must be converted before anything else happens with is. The program digtoxyz is one program for this conversion. It takes digitized datasets in a variety of formats and tries to identify what the contents are, then converts and saves them in the appropriate manner. The reason for the complexity lies in the diversity of things we digitize, from sock electrodes points to the tips and rods, and the need to treat each one a little differently. A further complexity lies in the two different devices we use to digitize; on good days this program handles both with grace and beauty.

Below we list each of the different types of geometrical objects or information that digtoxyz knows about and what type of input file format is necessary for consistent interpretation. I have removed all references to what used to be called the ``old digitizer'', this was the one built at the CVRTI that fixes the object on a rotating jig and moves the caliper to touch the points. We no longer use nor support this device. In contrast, the ``Microscribe digitizer'' is the device we have used for 10 years now. We create files from this device using Phil's program on the Macintosh. The standard extension for these files is ``.ms''.

2.1 Digitizing Tips

There is no magic nor strict paradigm for digitizing with the Microscribe; on the other hand, a few lessons from experience can be useful:

  1. Always try and perform the entire digitizing pass with the same settings and zero marks on the digitizer. Resetting the digitizer also re-zeros the origin and so locations digitized with one reference angle cannot be merged with those digitized with a different one. If this happens accidently, then try and at least re-digitize something in the new reference frame that has already been sampled before under the original reference and make sure to tell someone, ideally Rob, that this has occurred.
  2. Try and save output files from the digitizer with the extension .ms. This way, there is less ambiguity about the contents and how we need to read them.
  3. The largest source of true error in digitizing seems to be the movement of the heart during the digitization process; the Microscribe should improve this because the object can be kept stationary while the sampling probe moves around it. However, reducing movement remains the most straightforward means of improving accuracy.
  4. If in doubt, collect the data, convert it immediately, and then view the results with map3d3d. If still in doubt, digitize it a second time. If still in doubt, ask for help! It's OK to be in doubt, but silly not to do something about it.

2.1.1 Order of digitizing

The program digtoxyz does not explicitly care about the order of items digitized, nor which itens are included in a single .ms file. However, a little planning in this regard can save time in the subsequenct processing.

2.1.1.1 Sock Points:

In digitizing a set of points from, for example, an epicardial electrode sock, the order of entry of the points in the .ms file determines the order the point will have in the .pts file that digtoxyz produces. So order is important here!.

2.1.1.2 Coronaries:

At present these are also read and converted into landmarks in the order in which they are stored in the .ms file. The numbering scheme in the input file establishes the layout of the landmarks file so order is only semi-critical here.

2.1.1.3 Recording needles:

For digitizing recording needles (via the rods), digtoxyz keeps track of the rod/needle number used in the .ms file and saves this information in the .lmarks file that results. Thus it is feasible to digitize in an order that is spatially coherent or just convenient and then have digtoxyz sort things out. In the subsequent needle geometry processing steps described in this manual, the needle info file should also be organized according to the same needle numbering scheme as the .ms file, but need not have any particular order. The software takes care of matching according to needle/rod number as they are labelled in the .ms file.

Crucial in the entry of needle/rod points is the order in which the points of each needle/rod get digitized. The convention as we have agreed to it so far is outer end = rod tip, then inner end = rod entry point.

2.2 Running digtoxyz

To run the program digtoxyz requires the following command line:

        digtoxyz [-m -c] infilename outfilename

where

-m
= metric units on old digitizer
-c
= confused units on old digitizer (radius in inches, Z in decimeters)
infilename
= name of the input digitizer file.
outfilename
= name of the output file, whose type depend on the contents of the input file; there may even be several output files that result from a single input file.

Note: New digitizer includes the units in the file while the old digitizer defaults to inches for units.

2.3 Sock electrodes

Sock electrode locations are read as a series of points, and then later triangulated and formed into geometric meshes that form the basis for visualization, computer models and other types of data analysis. Hence the role of digtoxyz is fairly simple--convert each point from digitizer format to a geometry file (either .geom or .pts/.fac files).

2.3.1 Digitizer data format

Header Lines (4 in all)
FILE NAME: filename LAST MODIFIED: Date and Time
Free line of text
``Measurements are in'' millimeters or inches)
****************************************
File Body
``Point'' # X-value Y-value Z-value
Example
zero 0.000 0.000 0.000
sock 1 28.508 -173.895 221.950
sock 2 20.061 -178.593 225.062
sock 3 11.862 -179.161 237.611
sock 4 5.729 -175.170 251.998

2.3.2 Output files

The result of running digtoxyz on epicardial points is a file that is part of the CVRTI geometry files. Its extensions is .pts and it contains a list of all the points, converted into Cartesian coordinates and units of millimeters.

2.4 Coronaries

These are the files that result from digitizing the paths of the coronary arteries, which needs to be done according to the following tips:

2.4.1 Digitizer data format

There are two slightly different file layouts that digtoxyz will recognize as being from a coronary artery tree. One contains extra information of a radius for each point in the tree. This eventually determines the thickness of the tubes drawn in map3d and if fixed at the time of digitization, then it supplies one set of values that digtoxyz otherwise requires from the user when it runs.

The first format includes only the coordinates of each artery point.

Header Lines (4 in all)
FILE NAME: filename LAST MODIFIED: Date and Time
Free line of text
``Measurements are in'' millimeters or inches)
****************************************
File Body
``Point'' # X-value Y-value Z-value
Example
Artery 1-1 28.508 -173.895 221.950
Artery 1-2 -.9385 4.8480 225.062
Artery 1-3 -1.0225 4.5970 237.611
Artery 1-4 -1.1590 4.4565 251.998
Artery 1-5 -1.2280 4.2110 221.950
Artery 1-6 -1.2645 3.9330 225.062
Artery 1-7 -1.2510 3.6935 237.611
Artery 1-8 -1.2050 3.5000 251.998
Artery 2-1 -.9345 4.8945 221.950
Artery 2-2 -.8620 4.5170 225.062
Artery 2-3 -.7850 4.2745 237.611
Artery 2-4 -.6610 3.9005 251.998
Artery 2-5 -.5445 3.6085 221.950

The second format includes a radius value for each point, something one must typically add manually.

Header Lines (4 in all)
FILE NAME: filename LAST MODIFIED: Date and Time  
Free line of text  
``Measurements are in'' millimeters or inches)
****************************************
File Body
``Point'' # X-value Y-value Z-value Radius
Example
Artery 1-1 28.508 -173.895 221.950 .25
Artery 1-2 -.9385 4.8480 225.062 .15
Artery 1-3 -1.0225 4.5970 237.611 .14
Artery 1-4 -1.1590 4.4565 251.998 .14
Artery 1-5 -1.2280 4.2110 221.950 .13
Artery 1-6 -1.2645 3.9330 225.062 .1
Artery 1-7 -1.2510 3.6935 237.611 .05
Artery 1-8 -1.2050 3.5000 251.998 .01
Artery 2-1 -.9345 4.8945 221.950 .15
Artery 2-2 -.8620 4.5170 225.062 .12
Artery 2-3 -.7850 4.2745 237.611 .1
Artery 2-4 -.6610 3.9005 251.998 .05
Artery 2-5 -.5445 3.6085 221.950 .01

2.4.2 Output files

The output file for coronaries has to be more complicated than the simple list of points that results from the epicardial sock points. The file format used for this information is called a landmark file and the details of its purpose and format are in the manual for map3d. Briefly, this is another simple text file organize by landmark type and segments. There are several different landmark types, one of which is a coronary segment, that consists of a list of all the digitized points, ordered according to the numbers supplied in the digitizer file.

2.5 Recording Needles (via rods)

Needles are usually digitized by replacing each needle with a rod of known length and then digitizing the tips and entry points of each needle. From this information and knowledge of the needle electrode size and number of electrodes we can construct the locations of all the needle electrodes. Recently, we have used two points on the needle to determine direction, removing the need for rods to replace each needle. For historic and compatibilty reasons, however, the key word in the .ms file for both rods and needles is ``rod''. The order assumed for the points on each rod or needle is outer end = rod tip, then inner end = rod entry point.

The first step is to convert the entry points and tips into a landmarks file that becomes the input to both map3d and subsequent processing programs described in Section 3 of this document..

2.5.1 Digitizer data format

Header Lines (4 in all)
FILE NAME: filename LAST MODIFIED: Date and Time
Free line of text
Measurements are in millimeters (or inches)
****************************************
File Body
``Rod'' #-1/2 X-value Y-value Z-value
Example
Rod 1-1 28.508 -173.895 221.950
rod 1-2 20.061 -178.593 225.062
rod 2-1 11.862 -179.161 237.611
Rod 2-2 5.729 -175.170 251.998


Notes:

2.5.2 Output files

The output file from digtoxyz is a landmarks file, this time with objects of type rod. This file is already in the form that map3d uses so that quality control the digitization and conversion is quite easy. For example, with a points file from the epicardial surface called 24mar96_sock.pts and a rod landmarks file called 24mar96_rods.lmarks a sample command line for map3d would be:

       map3d -f 24mar96_sock -lm 24mar96-rods.lmarks

The main utility of the rods landmarks file, however, is to convert it into needle electrode positions, for which we have the program needlegeom , described in section 3 of this document.

2.6 Stimulus sites, occlusions, and other single locations

Most single locations are organized in a similar fashion in digtoxyz , as is the resulting landmarks file, which is the output format for such sites.

2.6.1 Digitizer data format

Header Lines (4 in all)
FILE NAME: filename LAST MODIFIED: Date and Time
Free line of text
Measurements are in millimeters (or inches)
****************************************
File Body
Point-type # X-value Y-value Z-value
Example
occl 1 28.508 -173.895 221.950
occl 2 20.061 -178.593 225.062
pace 1 11.862 -179.161 237.611
infusion 1 5.729 -175.170 251.998
lead 1 9.729 -134.170 123.998

Note the example contains names for each of the various single points that we currently recognize. ``occl'' = occlusion; ``pace'' = pacing site, ``infus'' = infusion site; and ``lead'' = selected lead. These items can appear as part of other digitizer files, eg., at the end of a file containing coronaries.

2.6.2 Output files

The output file is a landmarks file. A single landmarks file can contain any mix of the different objects we have described so far. Hence you may wish to digitize coronaries and single points on one file, then have a separate file for needles, and a third file for sock electrodes.


3 Landmarks to geometry: needlegeom

The purpose of needlegeom is to convert the contents of a landmarks file containing rod or needle endpoints into a .geom file containing needle electrode locations. This output file then serves as input for map3d and any program that reads needle geometry.

3.1 Using needlegeom

To perform this task, needlegeom needs some information, some of which is part of the command line, with the rest (optionall) coming from supplementary files. The command line arguments are as follows:

     needlegeom lmarkfile geomfile
          [-i needleinfofile] 
          [-n num_elecs_per_needle]
          [-e elec_spacing(mm)] 
          [-o offset] [-c channeloffset]
          [-t transmatrixfilename]

lmarkfile
is the name of the landmarks file that contains the rod information. This file is described in Section 2 of this document. The landmarks file can contain other information but only the rod entries are used.
geomfile
is the name of the output file you wish to make--a standard, CVRTI, .geom file
-i needleinfofile
is a file we shall name ``needle info file'' because it contains the links between needle numbers and channel numbers from the experiment (channel numbers), as well as the spacing and number of electrodes in each needles. The format of the file is described in section 3.2 below. This file is optional, but only if the channels for the needles all fall in sequence, with no gaps, and the spacing and number of electrodes is constant for all needles so that we can describe them completely by the following command arguments.
-n num_elecs_per_needle
is the number of electrodes per needle that you have used. The program assumes that this number is constant for all needles in an experiment--if this is not the case, it is necessary to specify a needle info file..
-e electrode_spacing
is the distance in millimeters between adjacent electrodes on the needle. The default value is 1.6 mm so this parameter is only necessary if the spacing differs from 1.6. Here too, we assume that the same value can be used for all electrode pairs--if this is not the case, a needle info file is necessary.
-o offset
is the distance in millimeters between the entry point on the rod and the first electrode, which has a default value of 0.5 mm. Only specify this option if the offset value is different from 0.5 mm. If it is not constant for all rodes/needles, then a needle info file is necessary.
-c channeloffset
is the offset applied to the first channel number assigned to a needle electrode. This is used only when the needle info file is not included in the command line. We assume in this case that electrodes follow in strict sequence for all needles, with no gaps in the order. So needle 1 would have a channel number of channeloffset + 1, then needle 2 would follow immediately behind in the ordering, and so on for all needles!!! This is probably the exception rather than the rule because of missing or broken needles but it is convenient and removes the need for an explicit needle info file.
-t transmatrixfilename
supplies the name of a file that contains the transformation matrix computed by align in a separate operation. See Section 3.3 for details.

3.1.1 Examples

If you have a needle info file for the experiment dated 24mar96, the command would be :

  needlegeom 24mar96-needles.lmarks 24mar96-needles.geom -i 24mar96.ni

We assume that the landmarks file 24mar96-needles.lmarks already exists and that the output file 24mar96-needles.geom does not (or you wish to overwrite it).
The file 24mar96.ni is the needle info file, which we describe next.

If we had a simple case in which all needles in the landmarks file followed in order, and there were 10 electrodes per needle, with otherwise default values the command line would become

  needlegeom 24mar96-needles.lmarks 24mar96-needles.geom -e 10
or if needles began at channel 193
  needlegeom 24mar96-needles.lmarks 24mar96-needles.geom -e 10 -o 192
Note that the value after the -o option is an offset applied to all channel numbers so for offset = 192, the first needle channel is 193.


3.2 The needle info file

A needle info file serves a number of purposes in telling needlegeom :

  1. the number label of each rod or needle,
  2. the data channel numbers associated with each needle.
  3. the number of electrodes in each needle,
  4. the spacing between electrodes along the needle, and
  5. the offset distance from the entry point to the first electrode of each needle.

3.2.0.1 Needle number label:

Each needle in the landmarks file can have a label associated with it that tell the program (and the human reader of the file) what ``needle number'' is associated with the entry in the file. This needle number is usually assigned according to some labeling scheme used during the experiement. As described in Section 2, this number also appears in the original digitizing file and is simply carried through to the landmarks file, and to the needle info file. The advantage of the label is that there is no explicit ordering of entries in either the landmarks or the needleinfo file, yet they will be aligned properly by needlegeom at execution time.

3.2.0.2 Channel numbers:

the same as the number of the input channel of the multiplexer. We assume that with each needle, the order of electrodes maps directly to an ascending series of channel numbers with the innermost electrode the lowest channel. So a single needle might run from channels 193-202, going from the tip of the needle back toward the wire.

3.2.0.3 Needles per electrode:

In principle, this parameter could take on any value but in practrice, the current needles have either 3, 10, or 12 electrodes and needlegeom checks againstthese values for ``sanity'' at run time.

3.2.0.4 Needle spacing:

This can be a single value or a set of three values, depending on whether the needle contains 10 or 12 electrodes. If we have 12 electrodes, two of them lie outside the heart and have a different spacing. The configurations are illustrated in Figure 1 and the distances in the figure have to be stored in the needle info file. For inside electrodes, we assume that the value is constant over the whole needle, the next two values indicate the spacing to the first outside electrode and between the first and second outside electrodes, respectively. For all spacing and offset distances, we assume units of millimeters.

Figure 1: Dimensions of the cardiac needle electrodes. Currently there are three physical configurations: a) 12-pole, spacing 1.0 mm (RV), b) 12-pole, spacing 1.6 mm (LV), and c) 3-pole, spacing 1.6 mm (most distal 3 of the 10 poles.) The 12-pole needles can, however, be connected so that only the 10 intramyocardial electrodes are available and included in the geometry.
 
needle diagram

3.2.0.5 Offset:

This is the distance between the entry point and the first inside electrode of the needle.

3.2.1 Needle info file format

The format of the needleinfo file is an ASCII, or plain text file, organized as follows:

Format of needle info file
Comment lines:
# A comment  
Header Line:
numneedles Number of needles in the file
Subsequent lines
First field Needle label number
Second field Channel number of distal tip electrode
Third field Number of electrodes per needle
Fourth field Spacing distances, in mm: interelectrode [outer_1 outer_2]
Fifth field Offset distance in mm to first electrode


Sample Needle Info File
# A sample needle info file for you
# It contains lots of variety
14 14 needles described in the file
1 321 10 1.2 0.5 First needle begins at channel 321, so runs from 321-330
2 331 10 1.2 .5 Second needle begins at channel 331, so runs from 331-340
3 341 10 1.2 0.5 There are 10 electrode per needle and 1.2 mm spacing
4 351 3 1.6 0.5 Only three electrodes in this needle
5 354 10 1.2 0.5 Note the jump in channel # of only three from the last needle
6 364 10 1.2 0.5  
8 384 10 1.2 0.5 Jump in needle label; # 7 fell out
9 394 10 1.2 0.5  
10 404 12 1.6 2.0 5.0 0.5 12-pole needle, spacing changes to 1.6 2.0 amd 5.0 mm
11 416 12 1.6 2.0 5.0    
12 428 12 1.6 2.0 5.0 0.5  
13 438 12 1.6 2.0 5.0 0.5 Channels jumps by only 10; last 2 electrodes of previous needle not wired
15 448 10 1.6 0.5 Skipped another needle label
16 458 10 1.6 0.5 14 needles but last one has label 16

To create these files, use any editor (if you use Microsoft Word, be sure to save the file as text only, without line-breaks) and then transfer the file to the SGI directory where the landmarks and geometry files live. The standard extension for the needleinfo file is .ni or .needleinfo. 1


3.3 Transformation matrix

The program to place epicardial sock geometry into the reference frame of the torso tank is called align. More generally, align will align any two sets of points according to some subset of points from both files that should be approximately the same. It generates a transformation matrix that any program can multiply with the nodes of the geometry to apply the same transformation to any other points. When we run align we apply the matrix automatically to the nodes of the geometry, and any landmarks in a file optionally included on the command line. Usually, we don't include the rod points in this landmarks file (although this would work too) yet we need to apply the same transformation to the needle points that we apply to the sock (and coronaries, stimulus sites, or any other landmarks that we digitize at the same time as the sock). Thus, the transformation matrix.

The format of the transformation is very simple, and the align program actually generates the text for this file, and writes it at the time of execution to a log file for your convenience. Just make sure the file exists, and the include it in the argument list for needlegeom .


4 Triangulation of needle geometry: needletri

4.1 What is needletri ?

The program needletri has a simple purpose--to take the triangles already entered for surfaces of a needle geometry and propagate them through a subset of the other surfaces in that geometry.

It is effectively the final stage in going from digitized rods to a needle geometry and other sections of this document describe the other steps.

4.2 Using needletri

The command line for needletri is as follows:

     needletri infilename outfilename
where infilename is the name of the input geometry file and outfilename the name of the desired output filename (which can be the same as the input).

All that is necessary is to have at least one surface in the input file triangulated and the program applies this same triangularization to any other surface in the file. The user controls which surfaces get propagated to which others and the program does some sanity checks for the number of points in each surface (i.e., you cannot take a triangulated surface with 10 points and propagate it to a surface with 12 points). Another assumption is that we have a regular array of needles with little or no crossing of needles within the tissue. But in the worst case, some manual editing of the resulting file will be necessary with map3d.


5 Extracting needle planes: makeplanes

The purpose of the makeplanes program is to extract sets of needles from a multi-needle geometry file and put them together in a separate file for display of data.

5.1 Using makeplanes

      makeplanes in.geom out.geom config.ni [-o] [-c]
where
      in.geom is the input geometry file
      out.geom is the desired output geometry file
      config.ni is the needle info file
      -o   = assign channels in .ni file to outermost electrode
      -c   = close the ends of the planes

In more detail

in.geom
is a regular .geom file in the CVRTI format. It should contain geometry for some number of needles from which you wish to extract a subset.
out.geom
is the output file that makeplanes will create for you (when all goes well).
config.ni
is the needleinfo file, described in Section 3.2 of this document.
-o
is an optional argument that tells makeplanes to assume that the channel number in the needleinfo file refers to the outermost electrode and not the innermost, as is the convention.
-c
an optional argument that causes makeplanesto lace the first and the last needles together.

Note that the input file will have multiple surfaces, one for each of the electrodes in the needles. The output file, on contrast, has just one surface and the triangulation of this file is across the needles, in a sense orthogonal to the triangulation of the input file.

5.1.1 Examples

   makeplanes bt11may94.geom  two24-1092.geom two24-1092.ni -o
The famous case which initiated the -o option. This one takes needles from the file bt11may94.geom according to the contents of two24-1092.ni and makes a new file called two24-1092.geom. The -o option ensures that the electrode order along the needle is reversed, i.e., that the channel numbers in two24-1092.ni refer to the outermost electrode.

The file two24-1092.ni looks like this

   9
   1 24 12 1.6 2.0 4.0 0.5
   2 192 12 1.6 2.0 4.0 0.5
   3 336 12 1.6 2.0 4.0 0.5
   4 408 12 1.6 2.0 4.0 0.5
   5 600 12 1.6 2.0 4.0 0.5
   6 504 12 1.6 2.0 4.0 0.5
   7 888 12 1.6 2.0 4.0 0.5
   8 1032 12 1.6 2.0 4.0 0.5
   9 1092 12 1.6 2.0 4.0 0.5

So 24 refers to channel 24, which is on the outer end of the needle (13 is on the inner end).

About this document ...

Programs for Needle Geometries

This document was generated using the LaTeX2HTML translator Version 2002-2-1 (1.71)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, 1999, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -split 3 -no_white -link 3 -no_navigation -no_math -html_version 3.2,math -show_section_numbers -local_icons needles

The translation was initiated by Rob Macleod on 2007-06-23


Rob Macleod 2007-06-23