phflot
commented
10 months ago Hello, sorry for the late reply, the code is currently only for 2D problems. It is possible to call it with individual stacks as channels, but then the output would be the x-y corrected stack. If your range of motion is limited, would could be possible is to run any simple 3D pre-alignment (e.g. cc based) and then use the flow registration approach for refinement of either individual slices or each pre-aligned stack over time.
Hello,
I'm having trouble understanding how to set up flow_registration for 3D volumes. We're collecting 1 stack every 30 minutes with a bright, sparse sample. It's pretty easy to see that the sample is drifting in x, y, and z.
Can you please provide some info on whether it's possible to coregister these samples across x, y, and z. For example, I'm confused on how to provide my data to the software. My guess is that each Z-level should be its own channel?