In the implementation I had started to write, a "Node" was simply a location in the MRF grid, while a patch was simply a region in the image. You could "apply" a patch to a node which meant to "place the patch at the node". Is this kind of the same thing that happens with your code?
You said "The Patch class contains the result of the Priority-BP solution, and is what get passed into the compositor instance. It's not used during the bulk of the solve process, but Labels are. They're almost synonymous, but have different uses depending where in the pipeline we are, in that Patches carry some extra data. Your confusion tells me that one or the other should be renamed."
Do these need to be renamed? Or can you clarify this a bit more? Or better, could these two classes actually be merged?
In the implementation I had started to write, a "Node" was simply a location in the MRF grid, while a patch was simply a region in the image. You could "apply" a patch to a node which meant to "place the patch at the node". Is this kind of the same thing that happens with your code?
You said "The Patch class contains the result of the Priority-BP solution, and is what get passed into the compositor instance. It's not used during the bulk of the solve process, but Labels are. They're almost synonymous, but have different uses depending where in the pipeline we are, in that Patches carry some extra data. Your confusion tells me that one or the other should be renamed."
Do these need to be renamed? Or can you clarify this a bit more? Or better, could these two classes actually be merged?