How can segmentation be made tighter? It's evident that in the example, the segmentation of the dog is not tight enough; I only want to segment the outline of the dog, without introducing any other background information.
2.Can the sd model used for filling be fine-tuned? I want to train a new category, such as "sfodglbns," representing a new and unseen species, and then use this species to fill in the tightly segmented outline.
Hi,I need some advice on this
2.Can the sd model used for filling be fine-tuned? I want to train a new category, such as "sfodglbns," representing a new and unseen species, and then use this species to fill in the tightly segmented outline.
thanks a lot!