Consider how best to handle "multi-channel" color spaces (RGB, HSB, CIELab, etc.)

@landinig wrote:

> The issue I see is that RGB are a special kind of multichannel that do not follow the generalisation of multichannel images (think of other colour spaces if implemented, like HSB or CIE Lab).
>
> Currently IJ1 does not display sliders unless it is a composite image (where you can process a particular channel). But all filters applied to an RGB image will process all the channels one after another. Not sure if changing this behavior might break things?
>
> Will IJ2 have RGB stacks? If so, colour images should not have a slider (because currently it does not do anything). The slider should appear after converting from RGB composite to RGB stack (like in IJ1).
>
> If there is no RBG stack type in IJ2, and colour images are generic stacks, then slider is not working, and filtering (blur, etc) need to be taken care of differently.
>
> Take for example "sharpening". On an RGB image this filter is applied to all channels, but in HSB or CIELab this should be applied only to the "V" or "b" channels. Not that I want to redefine filtering depending on the colour space, but there are new challenges in treating colour spaces as generic multichannel.

Migrated-From: http://trac.imagej.net/ticket/1108

The issue names no files, tests, or entry points. Start by reviewing the current handling of RGB and generic multichannel images, then clarify the supported color-space behavior and agree on a concrete scope before implementation; completion is not defined in the issue.

Written by the indexing model from the issue text.