[IPOL discuss] optic flow algorithms in ipol

[eml]

> Should we make one publication for the Horn & Schunck algorithm with
> many sub-parts devoted to each implementation and links to many demos?
> Should we compare the implementations internally and pick one (or
> two)?
> What do you think?

The paper of Horn & Schunck proposes an implementation in full detail.
 Any two correct implementations of this paper will produce exactly
the same results, so there is no point in having more than one
version.

Otherwise, an unpublished workshop comparing many implementations is
interesting as a tool.  Indeed, all the implementations of this
algorithm that I tried, including my own, produce vastly different
results.