Information regarding MIRAX v2.2 tile positioning

[John Minnie]

Hi List,

 

For the past few months I have been working on bug #92 on the openslide
issue list. The main problem is that the new format (greater than v1.9)
appears to have missing position data, which manifests as seams when
positioning the tiles. It is my contention that the "usual" slide position
file present in v1.9 data has been dropped, in favor of an xml-based layout
of the same positioning data.

 

In the hierarchical section of the ini file, one will find several xml
related NONHIER sections. I have been able to consistently extract from the
data files some xml-based information that seems to be 1-to-1 related to the
slide position file of old. I have to get final confirmation, but I think
the xml-based position data is found in the "ProfileXML" section of the
"Scan data layer" NONHIER level. 

 

I have a zip archive of 1.5Mb with both sets of data. If you are interested
in reviewing the data, please let me know then I can forward it directly to
you. In the archive you will find a dump of the original slide position
buffer, where the *.xyp files are named as follows: {slide position record
number}_{x-position}_{y-position}.xyp. The slide in question from which this
data was extracted is the CMU-1 test slide available on the openslide site.

 

The xml scaninfo files are also present, and the names of these files are
the tile index to which the scaninfo applies, extracted directly from a data
file and not computed based on any formula.

 

For all of the v1.9 MIRAX slides at my disposal, there is ALWAYS a 1-to-1
correlation between the total number of xyp files and xml files, supporting
the notion of a relationship between the old slide position file buffer
previously found in the NONHIER section "VIMSLIDE_POSITION_BUFFER" and the
xml-based scaninfo data. For v2.2 slides, the old position buffer cannot be
located, but the xml-based data is always present.

 

My quest is now to seek consensus from the openslide community that this is
a real correlation and not coincidence. It is my further contention that the
"VIMSLIDE_POSITION_BUFFER" data has been removed from v2.2 slides (and
possibly in v2.0) due to redundancy, and that the xml-based positioning is
the only available data for tile positioning.

 

Any advice and suggestion how to interpret the xml-based position data in
terms of the old slide position buffer will be hugely appreciated. My own
work in this area continues.

 

Regards and thanks,

John

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.andrew.cmu.edu/pipermail/openslide-users/attachments/20120911/d2106f78/attachment.html