Re: Whole slide image analysis with ImageJ

Posted by dscho on Apr 19, 2011; 11:43pm
URL: http://imagej.273.s1.nabble.com/Whole-slide-image-analysis-with-ImageJ-tp3684908p3684912.html

Hi,

On Tue, 19 Apr 2011, Aivar Grislis wrote:

> For the ImageJDev project we are using Johannes' Fiji ImgLib as our
> underlying data model.

To clarify: ImgLib was conceived and developed by Stephan Saalfeld and
Stephan Preibisch since spring 2008.

Recently, other developers (I only play a minor role, personally)
contributed to, and pushed forward, ImgLib, including Curtis Rueden,
Tobias Pietzsch, Barry DeZonia, Lee Kamentsky and others.

> The usual ArrayContainer is limited to around 2 gigapixel images due to
> Java's use of integer array indices.

... but of course the ImgLib provides other means to store images which
are not limited to 2^31 pixels...

> There is a CellContainer that can be used to hold huge images that fit
> in memory or could support image tiles that are swapped out.

Do not forget the PlanarContainer, which is currently limited to 2^31
planes that can contain up to 2^31 pixels.

The ImgLib is able to accomodate more container types (without requiring
any change in the implementation of algorithms!) than Array, Cell and
Planar. It is even possible to implement your own, custom, container,
again without requiring to change _any_ implemention of an algorithm.

In my opinion it would be in the best interest of the authors of the
previously mentioned paper to develop their code on top of the ImgLib
since it is clearly well-adapted to extreme cases like huge planes.

Of course, I also think that it would be the best to publish the plugins,
since nothing is more useless than a program (next to) nobody uses.

Ciao,
Johannes