The idea is that we could calculate some velocity and acceleration from the flow-skeletons (created by artists using feature in issue #7) of a few previously displayed images and create a guess of where the flow skeleton should be in the next frame and hand that as a constraint for the finding of the next image.
The idea is that we could calculate some velocity and acceleration from the flow-skeletons (created by artists using feature in issue #7) of a few previously displayed images and create a guess of where the flow skeleton should be in the next frame and hand that as a constraint for the finding of the next image.