Possible to fit with only shape model

[Junkern]

I have been looking through the menpofit documentation. And, apparently, all the different fitters combine different models (e.g. for the AAM: shape model, motion model, appearance model). So, they are not only fitting based on the shape, but also on the color? Am I correct in assuming that?

So, my question is: Is there a fitter only working on the shape model, hence only taking the shape into account? Is it possible to work only with shapes, but not with textures?

Many thanks!

[patricksnape]

Just out of interest, how would you expect that to work? How would you fit an image without inspecting the pixel values?

[Junkern]

I think I expressed myself a bit wrong.
I cannot fit without inspecting the pixel values, you are right about that.

However, for the training of the model, I do not want to take the appearance into account, only the landmarks from the .pts file.
(If I spin this further, the cost function for the fitting then would not take the appearance into account, because the model does not have an appearance.) I hope this makes sense :D

[nontas]

I am still a bit confused. Could you please give a specific example of such cost function?

[patricksnape]

Even if you have a shape model, you need some cost function that minimises some error? What error can you minimise/probability can you maximise that involves an image but doesn't take into account it's appearance? Maybe if you describe what you want to use this for we can better help you formulate your desired cost function.