I would like to train a model that would be able to return an svg path for a given image, as in the image below:
Raw image:
Mask/label:
Is it possible to train such a model, and if so, how should I go about it? I've read about Image Segmentation, but I'm not sure if this is the best way to approach this problem, because when reading about this method, I found examples that only detect objects (fully connected shapes) and not lines.
I was wondering in which exact way I should specify "labels" for this task.
Ultimately, I would also like the model to be able to detect more than one such mask for a given image.
I've read about image segmentation, classification and other methods, but I'm not sure which one would be suitable for this problem. So far I have only dealt with the usual classification of images, but this problem seems to me to require a completely different approach.
