HOME GUIDE OPERATIONS DOCS ERRORS FORMATS INSTALL NEW TIPS WEB SITES

CA SI - Correspondence AnalysiS - Initialize

(08/22/97)

PURPOSE

First part of CORAN or PCA: Creates sequential file from image set, to be used in Correspondence Analysis - CORAN, or Principal Component Analysis - PCA, ('CA S'). This sequential file contains:
1. all image data within the mask array,
2. the mask address array,
3. the image identifications.

SEE ALSO

CA MAP [Correspondence Analysis - nonlinear mapping]
CA S [Correspondence Analysis - find solution]
CA SM [Correspondence Analysis - factor map]
CA SME [Correspondence Analysis - factor map and Eigen Values]
CA SR [Correspondence Analysis - full reconstitution of images]
CA SRD [Correspondence AnalysiS - reconstitute differential images]
CA SRI [Correspondence AnalysiS - reconstitute importance images]
CA SRP [Correspondence Analysis - PCA reconstitution]
CA VIS [Correspondence Analysis - Create visual map]

USAGE

.OPERATION: CA SI

.ENTER IMAGE FILE TEMPLATE : SEC***
[Enter a file name template identifying the image series to be analyzed.]

.ENTER IMAGE NUMBERS: 1-40,45,50-70
[Self-explanatory. If the data are supplied as a 3D stack, then the numbers are interpreted as slice numbers. If the 'R' (Random) option is used (see below), then the total number of images to be created is determined by the sequence entered here, although only the first file need be supplied.]

.MASK FILE: MAS002
[Enter the name of the file containing the mask. Only image pixels where mask pixels are greater than 0.5 are analyzed in CORAN and PCA. The mask address array is stored as part of the sequential file. It allows the user to recover the original [masked] images or to reconstitute images using selected factors (see also 'CA SR'). Current limitation for total number of pixels in mask: 4000.]

.SEQUENCE FILE: SEQ001
[Enter the name of the sequential file where data are to be stored.]

.ALL FILES EXIST (Y/N), (I)NV, (R)ANDOM: Y
[Options are read in as string; e.g., 'YI' or 'NR'.
If all files in the sequence specified exist, enter 'Y'.
If there are any gaps, enter 'N'.
The 'I' option is used to invert the contrast of the images as they are being read in.
The 'R' option is used to create an image set by successive random permutations, starting with the first image. Only the first image of the sequence needs to exist; it is the only one used in this computation.]

If 'R' is specified, the system asks an additional question:

.SHUFFLE FACTOR: 1.5
[Randomization is achieved by specification of a shuffle factor > 1. The total number of permutations in going from one image to its successor is equal to the number of pixels in the mask times the shuffle factor. A shuffle factor of 1 is not normally sufficient to eliminate correlation in the X direction. Suggested value is 2.]

.ADDITIVE CONSTANT: 0.0
[If your images have negative values, this can be corrected by specifying an additive constant that will be added to all (within-mask) pixels of all images]. The smallest additive constant needed to achieve positivity will be printed. Unlike CORAN, PCA does accept negative pixel values (in images); thus, it is not necessary to add a constant (although it will be displayed).

NOTES

  1. The next operation in the Correspondence Analysis (and PCA) sequence is 'CA S', the actual run of the factor analysis.

SUBROUTINES

JPMSK1

CALLER

UTIL1

© Copyright Notice /       Enquiries: spider@wadsworth.org