Tony Tung

`num_peaks_per_label` shouldn't affect the number of spots found. it _will_ affect the number of pixels found.

the footprint argument theoretically allows us to do what we want to do. It allows us to control the regions being considered for each axis. Unfortunately, it has an off-by-one...

Are there other edge effects that arise from padding with zeros? The biggest challenge here is probably that we'll be doing a _lot_ of copying.

I think we should do this ASAP.

We definitely need this; unless this has been superseded by a new issue, we should get this done ASAP.

@nickeener it looks like this is passing now. Ideally, you should just copy my changes into your PR and land them in the context of your PR. @berl any concerns...

++ zeroing out.

@shanaxel42 isn't this done?

This is potentially quite expensive (and destructive to our travis caches). We should stand up an environment with the caches disabled and wire it up to @joshmoore's periodic tests.

Sorry, not an environment, but just the setup you have for checking requirements.