Fitting curve and split it equally

Hi,

I have the attached image and I want to fit a curve between each two marked point and split each boundary eqaully. I know that one of the method is using angles and segment each boundary based on angles but I have to use curve fitting for this purpose. For instance, for this image, I have to split the two top and bottom boundaries to 10 and left and right to 16. I appreciate it if you could help me with that.

[Image Analyst]

I have no idea what you want. What does "split the two top and bottom boundaries to 10" mean? Ten what? Ten segments of boundary? You already have the boundary - a list of (x,y) coordinates I presume. So how did you get, or how do you want to get, the red spots?

[Bahareh Salafian Esfahani]

I need these e.g. 10 segments for TFI purpose. Yes 10 segments of boundary. Yes I have the coordinates of those white boundaries and I have 9 other images similar to the attached one and I want to have same segments for all of them, I mean for all, 10 for top and bottom and 16 for left and right. Those red marks are my landmarks and they are available in all of those 9 images. Is it clear now?

[Image Analyst]

No, it's not. So do you have a list of all (x,y) or (row, column) coordinates for the red spots already? Are you certain they land on the white line? If so, why can't you just find where (what indexes) in the list of white coordinates do the red ones live, and use interp1() to find 10 points in between to split them? It will get a little tricky for the list at the top and the bottom because you have to "wrap around" but you can handle that by doubling the coordinate list and using mod() or rem().

You might want to check out John D'Errico's interparc: interparc in his File Exchange (a virtual treasure chest of gems)