On 3/16/2013 2:27 AM, Ray Koopman wrote: > On Mar 14, 9:36 pm, Math Guy <M...@Guy.com> wrote: >> Ray Koopman wrote: >>> >>> You haven't said whether the ordering of the points around the >>> loop is given or not. >> >> The user will specify the markers to be used (there are other markers >> on the heart which will not be involved in this ring). A routine will >> identify the order by picking one point and then determine which of >> the remaining points is closest, and will piece-wise work it's way >> around the ring. This should avoid any errors if the user gives the >> order incorrectly. >> >>> Also, what do you want to do if the loop is not convex? >> >> Do you mean what if the loop looks like a potato chip with a bite taken >> out of it? >> >> I don't see how that would affect the proposed area calculation method. > > It won't affect the plane of best fit, but it may make it harder to > identify the loop. I would give the 2D coordinates to a convex hull > routine. If the hull uses all the points then it's your loop. But > if some points are inside the hull then a simple nearest-neighbor > approach can sometimes connect the points in the wrong order. >
When I had been trying to visualize the stated problem, I assumed that a "ring" would be a convex star.