That the last point of a curve should be fixed is not really a necessity of the semilandmark algebra. This only convenient because then it is easy to estimate the tangents along which the semilandmarks are allowed to slide by taking the vector of the two neighbouring points.
Instead of treating the last point as a landmark (which it isn't) I suggest to estimate a tangent for the last point and let it slide as well. best wishes philipp Am 16.06.2005 um 13:08 schrieb morphmet: > Hi All, > > I have a set of specimens with a curve which has one end obscured by a > mobile feature, so that end cannot be used as a landmark. > > I have set semilandmarks along this curve using a comb between one > landmark on the curve and another off of the curve. > > Clearly I can't use the off-curve landmark as the endpoint of the > curve, > so I am using the last semilandmark instead. > > My question: Is it valid to fix the terminal semilandmark, and define > the rest as sliders, or must they all be fixed? > > Many thanks, Shaun * Mag. Philipp Gunz Department of Anthropology University of Vienna Althanstrasse 14, A-1091 Vienna, Austria +43 1 4277 54729 -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
