Active contour algorithm and blurred edge.

I'm working on an algorithm which aim is to semi-automatically track the lumen diameter of some arteries (for example the common carotid artery) on sequence of US images.

The activecontour function (snakes) works well on images with high signal to noise ratio (SNR).

However, the snake misses boundaries when it comes to images with low SNR like the one as an attached file which has blurred edge on the right side of the artery.

I've read papers on shape correlation ("shape prior") techniques using low rank matrix approximation that create more robust algorithm.

The aim of the study is not to acquire images of better quality but to find a algorithm with better robustness.

However, even if i understand the general idea of shape correlation in active contour model algorithm, i am only starting on matlab and have hard time trying to compute this concept.

how can i make the activecontour function more robust ? which other function can i use?

Thanks for your help.