CZI: questions about openslide grids
mathieu.malaterre at gmail.com
Wed Dec 17 03:15:53 EST 2014
On Wed, Dec 17, 2014 at 8:43 AM, Benjamin Gilbert <bgilbert at cs.cmu.edu> wrote:
> On 12/16/2014 05:23 AM, SOUEDET Nicolas 215310 wrote:
>> Is there any other format supported that had to deal with this kind
>> of issue ? I guess this is not particular to Zeiss, but maybe I am
>> wrong ?
> The other formats that use irregular tile placements (MIRAX, Trestle,
> Ventana) record both a tile's actual pixel coordinates and its idealized
> placement on a regular grid. This allows the tilemap grid to do a "poor
> man's range search", using the ideal position as a starting point for a
> limited brute-force search.
> If CZI only records the pixel coordinates of each tile, we've reached
> the third option in my previous email. I don't think it's worthwhile
> trying to cleverly compute idealized tile advances (as Mathieu proposed
> on StackExchange) just so you can use the tilemap grid.
Just for later reference: I did not find the `idealized` position,
does not mean they are not stored somewhere. For instance I found the
jitter 10/15% somewhere in the XML file, which AFAIK is not documented
in the CZI reference I downloaded...
But I agree that a brute force approach should just work.
More information about the openslide-users