[ome-devel] Multi-spectral imaging formats

[Munro, Ian]

Hi Lee

I’ve had a similar issue with FLIM (Fluorescence Lifetime) data where up to 1024 time-points per pixel are stored contiguously in  a file.
The approach that I took was to buffer the whole file in memory  & reorder the first imd that a plane is requested.
Subsequent planes can then be returned from the memory buffer fairly quickly.

If you expect all 63 channels to be requested together more often than not then you could take the same approach.

Regards

Ian

On 26 Aug 2014, at 17:48, Lee Kamentsky <leek at broadinstitute.org> wrote:

> Hi all,
> I've been working on the .IM3 format for the Perkin-Elmer Nuance multi-spectral microscope. It has 63 channels per pixel for the spectral data and the data is interleaved (and I thought it would be the most efficient to keep it that way). For the IM3 reader, my naive idea was to package all 63 channels as one interleaved image, but I have a feeling that's 60 more than most software expects, but it would be very sad to run through the 200Mb file 63 times in order to get 63 planar images.
> 
> Otherwise, it looks like an awesome machine. I am guessing that LOCI has been working with other multispectral imaging systems and might point me to a representative reader that I might use as a guide.
> 
> --Lee
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel