Rafael Toledo

this change gets to work to me, @zhyma , thanks. but i'd like to know how `not isinstance(s_prev, list)` is equal to `s_prev == None` ? because, i'm debugging the...

The problem seems that you can't compare an array to a single None. To solve that i compare the class of the object to Nonetype `if s_prev.__class__ == 'NoneType'`, but...

Using only the mask logits did not work for me, it rendered nonsense results. On the other hand, querying positive and negative points from the binary mask yielded a better...

Thank you for your reply @thomasfermi. Yes, I converted all (u, v) coords to XYZ. I see that the only correct XYZ coords would be the ones that the (u,v)...

Sure @praeclarumjj3 , that's the image. 

Hey @jaime-1998 , just wanted to check in. Did you have any updates? Curious if you tried training with the unlabeled data or found a workaround. Thanks!

IMO, I'd not expect that dinov2 features work for fingerprint ID. Their feature are not "unique" in the sense that you want. The dinov2 features are designed to resemble the...

You could try creating a head or a projector over the dinov2 features that accomplish what you want, such as a siamese network schema with contrastive learning.

In your case, it seems that the object pose will not change too much, so dino features may help you, but I don't know how robust it can be for...