Segment the array prior to searching

[Matthew Whiting]

Since we are now working with a single set of stats for the entire cube, it is possible to segment the cube so that it is represented as a set of bools (ie. true/false) as to whether a given pixel is a detection or not.

This would then make the searching a more trivial exercise, since it will be a simple binary choice. You would not have to call the isDetection functions on each step, just look at the segmented array. There would need to be a new array created (albeit just of bools), but this could be done as part of a searching function (one of two ways, depending on reconstruction state) and then a generic cubic searching function could take over.

[anonymous]

Note that you will still need to search in both image and spectrum senses, unless you drop any minimum-size criteria, in which case you could just search eg. only in each spectrum.

[Matthew Whiting]

This is probably unnecessary now -- the way the searching is done, combined with the isDetection functions, means that we don't really need to create another full array.

I'm going to set this to "invalid" and hopefully ignore it from now on.