I’m implementing a custom stereo algorithm needs to use similar concept of “AggregateCostScanlines”. However my input is not a U8 vx_image but a cost analyzing lambda function with multiple floats as inputs. Can anyone share the algorithm of what this “AggregateCostScanlines” is?
We don’t release the source code for this function.
You can check OpenCV as alternative:
Actually, I’m looking for an GPU implementation. But thanks anyway.