Documentation
Data formats
Data should be uploaded either as a tpsDig (.tps) or .dat file. tpsDig files (which are easy to collect using Jim Rohlf's tpsDig and tpsUtil) are are of the format:
LM=0
OUTLINES=1
POINTS=200
61 207
58 216
...
64 197
61 206
IMAGE=176 Acer neg01.tif
ID=1
LM=0
...
If the first line of your dataset starts with "LM=", it is read as a tpsDig file. Specimen names are read off the IMAGE line (generated automatically by tpsUtil). Your specimen names can't contain "LM=", "IMAGE=", or "SCALE=".
Alternatively, .dat files are of the format:
OBJECT Vulpes 001
SCALE 0.014792899
1452 1012
1436.4 1733
...
2882.8 955.3
2870.1 972.8
LANDMARKS
2692 1032
1948 1104
OBJECT Vulpes 002
...
The delimiter "OBJECT" is key (case is ignored). Specimen names are taken from the line with "Object". Your specimen names cannot contain "landmarks", "LM=", "scale" or "object". The section with landmarks comes at the end of each object (you may alternatively use the tpsDig style "LM="). Standard eigenshape doesn't use the landmarks, but it is okay with them being present in the file (they are ignored).
Either format can have a line with SCALE or SCALE=, if you wish to include a scale factor. The data files may have empty lines, but may not have a title line. There may be problems with Mac text format files. If you discover the best way prepare these files, let me know. I find that, on a PC, saving as text from Microsoft Word generates good text files. I edit the text files in Notepad.
|