writeData: Write Datasets in different Formats
In PET: Simulation and Reconstruction of PET Images

Description Usage Arguments Details Value Author(s) References See Also Examples

The function offers the possibility to write ".txt", ".dat", ".dat.gz", ".pet" or ".fif" files. Furthermore the graphic formats ".tif", ".tiff", ".pgm", ".ppm", ".png", ".pnm", ".gif", ".jpg" and ".jpeg" are supported. The function was especially written to handle sinograms and datasets, that are processed with the functions radon, iradon and iradonIT of the R package 'PET'.

1 2	writeData(data, outputfile, fileHeader = NULL, imType = "normal", fileOverwrite = "ASK", DebugLevel = "Normal")

`data`	(matrix) The data to be written out.
`outputfile`	(character) `outputfile` naming the file to write to, including the pathname. The path has to be given relatively to the working-directory of your R-session or it contains the full path of the file. The extension of the filename specifies the format. Currently, ".txt", ".dat", ".dat.gz", ".fif", ".pet" and the graphic formats ".tif", ".tiff", ".pgm", ".ppm", ".png", ".pnm", ".gif", ".jpg" and ".jpeg" are supported. See below to details to get more information about these formats.
`fileHeader`	(list) If the file format "is .dat", ".dat.gz", ".pet" or ".fif", then `fileHeader` is necessary. If in these cases `fileHeader=NULL`, default values will be generated dependent on `imType`. See below to details to get more information.
`imType`	(character) `imType` specifies the type of `data` and thus the default values of `fileHeader`. Default is `imType="normal"`. Additional implementation is `imType="radon"`, i.e. that the image is a sinogram (Radon transformed image).
`fileOverwrite`	(character) Control the behaviour for overwriting a file. Supported are "NO", "YES" and "ASK". If fileOverwrite = "ASK" the routine will ask you in case of an existing outputfile with this name. Default is "ASK".
`DebugLevel`	(character) This parameter controls the level of output. Defaults to `DebugLevel="Normal"` for a standard level output. Alternative implementations are `"Detail"` if it is desirable to show almost all output on screen or `"HardCore"` for no information at all.

In the following different supported formats of inputfile are explained. For ".dat", ".dat.gz", ".pet" and ".fif" the default values dependent on imType are defined.

".tif", ".tiff", ".pgm", ".ppm", ".png", ".pnm", ".gif", ".jpg" and ".jpeg"
For using these graphic formats the R package adimpro is necessary. See there, to get more information about these formats.

".txt"
Writes files in ASCII format without header, which contain only the raw data. The R-routine write.table is used.

".dat" and ".gz"
Write files in ASCII format where files contain a header with three or four rows followed by the data. Compared to ".txt" the advantage is that is possible to save and load higher-dimensional data. Furthermore important values are saved, that are needed for Radon and the inverse Radon transformation. If the extension ".gz" is specified, then files in ".dat" format will be stored which are additionally compressed by 'gzip'.
Assuming the data are two-dimensional, at first fileHeader will be written out. Default values will be defined (dependent on imType) for not specified parameters. The file has only the following form:

Description: char (optional)
SignalDim: int int
XYmin: double double
DeltaXY: double double
... and after the header data will be written out

Explanation of the parameters:

Description:	Can be a short description of the data. It is possible
	to leave out this parameter. In this case the first
	line of the file begins with 'SignalDim'.
SignalDim:	In case of 2-dim data, 'SignalDim' is the number of
	rows and columns of the array, but also a greater
	dimension is permissible.
XYmin:	Leftmost coordinate and lowest coordinate of the
	data.
DeltaXY:	Quantization steps in image, in x and y-direction.

Defaults to:

	`imType="normal"`	`imType="radon"`

SignalDim:	`dim(data)`	`dim(data)`
XYmin:	`-0.5*(dim(data)-1)`	`c(0,-0.5*(ncol(data)-1))`
DeltaXY:	`c(1,1)`	`c(pi/(nrow(data)),1)`

If the dimension of data is greater than two, the number of 'SignalDim', 'XYmin' and 'DeltaXY' should be increased appropriately. Note in this case you must define the fileHeader.

".pet"
Writes a picture in ".pet" format (raw float with header) to the file e.g. inputfile = "FileName.pet" (binary format). The ".pet"-file has to be of the following structure: The first part contains a header and the second part the data, e.g. an image. Only the values of the parameters are saved in the header and not the parameter names.

Description:	(80-byte character) Should be a short description of
	image. Defaults to `Description="}\eqn{\backslash 0}{\0}\code{"`.
XSamples:	(integer) Number of rows of array. Defaults
	to `XSamples=nrow(data)`.
YSamples:	(integer) Number of columns of array. Defaults
	to `YSamples=ncol(data)`.
Xmin:	(float) Leftmost coordinate in original image.
	If `imType="normal"` then defaults to `Xmin=`
	`-0.5*(nrow(data)-1)` or if `imType="radon"`
	then `Xmin=0`.
Ymin:	(float) Lowest coordinate in original image.
	Defaults to `Ymin=-0.5*(ncol(data)-1)`.
DeltaX:	(float) Quantization steps in original image (x).
	If `imType="normal"` then defaults to `DeltaX=1`
	or if `imType="radon"` then `DeltaX=pi/(nrow(data))`.
DeltaY:	(float) Quantization steps in original image (y).
	Defaults to `DeltaY=1`.

The dimension of the data should be not greater than two.
The ".pet" and the sequencing ".fif" format will be used in the iradonIT method to store iteration-steps and to read a references image.

".fif"
Writes a picture in ".fif" format (raw float with header) to the file e.g. inputfile = "FileName.fif" (binary format). The ".fif"-file has to be of the following structure: The first part contains a header and the second part the data, e.g. an image. At first the values of the following parameters are saved in the file:

FIFIdType:	(integer) ID used to restore FIF:17737:'\0''\0''E''I'.
	Defaults to `FIFIdType=17737`.
FileName:	(100-byte character) Name used for saving this image.
	Defaults to `FileName=outputfile`.
Description:	See above.
Date:	(10-byte character) Date (YYYY-MM-DD). Defaults to
	`Date=as.character(Sys.Date())`.
XSamples:	See above.
YSamples:	See above.
ArrayType:	(integer) Defines the format number: 1 for Real or
	2 for Complex. Complex matrices are determined by A=a_{i,j}
	where a_{i,2n} is the imaginary part to the real value
	a_{i,2n-1}, i=1,…,M, j=1,…,2N, n=1,…,N.
	Defaults to `ArrayType=1`.
Xmin:	(float) See above.
Ymin:	(float) See above.
DeltaX:	(float) See above.
DeltaY:	(float) See above.
SignalMin:	(float) Lowest signalvalue in array. Defaults
	to `SignalMax=max(data)`.
SignalMax:	(float) Highest signalvalue in array. Defaults
	to `SignalMax=max(data)`.

After this header data are following.

writeData creates a file, but returns nothing.

Joern Schulz, jschulz78@web.de.

Schulz, Joern, Diploma Thesis: Analyse von PET Daten unter Einsatz adaptiver Glaettungsverfahren, Humboldt-Universitaet zu Berlin, Institut fuer Mathematik, 2006.

readData, radon

## Not run: 
P <- phantom(addIm="blurred1")
rP <- radon(P, RhoSamples=401)
irP <- iradon(rP$rData, 257, 257)
# Saving 'P' as a 'jpeg'-file
writeData(P, "Phantom.jpeg")
# Saving 'rP' as a 'pet'-file
writeData(rP$rData, "RadonPhantom.pet", fileHeader=rP$Header)
# Saving 'irP' as a 'dat.gz'-file'
writeData(irP$irData, "RecPhantom.dat.gz", fileHeader=irP$Header)
rm(irP,P,rP)

Image1 <- readData("Phantom.jpeg")
Image2 <- readData("RadonPhantom.pet")
Image3 <- readData("RecPhantom.dat.gz")
viewData(list(Image1,Image2$Signal,Image3$Signal))
rm(Image1,Image2,Image3)

## End(Not run)