Peter, Checkpoint does not currently support equidistant spaced semi landmarks along a multi-section curve. There's a couple reasons for this. The curves in Checkpoint are designed so that you decompose "long" curves into sections via the control points. The semis are distributed between the control point from section to section to ensure a consistent number of semis within each section. As such, a short section will have a higher density of semis than a longer section; this is likely the behavior that is causing you problems.

The curves are designed this way so that regions can be decomposed into sections that follow the natural curvatures present in your specimen population. It is usually possible to examine a subset of your population to identify inflection points along your curve that are identifiable across all specimen. You should place the control points at these inflection points (on all specimen), have the same number of curve sections on all specimen, and you'll end up with a consistent decomposition across your population.

Distributing points in this way better captures variation that may occur among the sections of the curves. For example, one section on one species may be short and that same curve section on another species may be rather long. And yes, in this case, the control points play the elevated status of a landmark point since the semi landmarks are spread between those.

The idea of equidistant spacing of semi landmarks along long curves comes from microscribe based techniques where there is no way to easily denote inflection points during collection. The most practical thing to do with a microscribe is to sample points as equidistantly as possible along the curve. Then, typically the equidistantly spaced semi landmarks are "slid" along the curve to reduce either the TPS energy or Procrustes Distance among your population. So, you don't really end up with equidistantly spaced points anyway.

https://www.stratovan.com/products/checkpoint?utm_source=rg&utm_medium=ans&utm_campaign=lm8

Thank you very much for your detailed answer David :)

Hi Peter,

I strongly suggest the use of Morpho R package.

#load example data

data(nose)

#control points

x

Peter Pogoda

If you use R the function ev.sp.points in Morpho could help. In addition, if you use Amira/Avizo the function B-spline allows you set the number of semilandmarks which will be positioned relatively evenly spaced. In any case during the sliding process the semilandmarks will be evenly spaced along the curves.

With Morpho it is easy achievable. Just needs a few command lines. Would recommend it to everybody digitizing curves in 3D space ;-) Thanks to Antonio!

If your "data" is already in R as an array including only curve points in 3D:

nspecs