My prediction looking at that example, is that the answer is NO, that there is no automatic way to extract that region without extracting any of the noise -- not unless those red boundaries are part of the original image.
The noise appears to be of different sizes and different curvatures, and we can see that it can run fairly close to the region of interest. My prediction is that for at least some of your images, the noise touches the region of interest in a way that does not permit completely accurate separation of the two.
However, it is not impossible that your noise is better behaved than I predict, so I would not rule out the possibility of being able to do what you want accurately
