VVVV.Packs.Image
VVVV.Packs.Image copied to clipboard
SolvePnP, CalibrateCamera: Coordinates VVVV cs. OpenCV
both SolvePnP and CalibrateCamera featuer an input called "Coordinates" i'd suggest to get rid of.
rationale: this is inside vvvv. so inputs and outputs of those nodes should be in vvvv-coordinates. if at some point you have inputs or need outputs in another format there should be converters that are not part of the nodes.
in the past (and still atm) those don't seem to work anyway. it seems setting to 'OpenCV' makes most sense atm. this takes inputs as they are and return results without conversion and it seems the results are correct.
thoughts?
this is mostly because i still sometimes get erratic results, and need to hack around the transforms. I agree that ultimately we just want VVVV style in/outs on everything