Can this be used for interactive segmentation?

[Zarxrax]

Can this model be used together with user input, for example where a user makes some clicks on an image to define what part of the image should be masked? Or does that require an entirely different kind of model?