multi-time-point multi-channel issues about affinder

[jni]

Yeah, I expect that this would work in the far future. Indeed the idea was for affinder to be able to align 3D volumes, for which you need four points before you start. I guess based on what you're saying, one of your images had 3 channels, and the other one was maybe 2D only? That would indeed cause a crash.

For now, affinder is assuming that all the dimensions of both images are spatial, and that both images have the same shape. We can probably add some UI elements to select which axes to align, and then affinder can create points layers of the correct dimensionality. Having said that, it will only work with trailing axes until napari is smarter with its axis matching.