What i dont understand is that by now why dont we have the ability to have it guess at at least some of the backs of the objects, or to separate the subject from the background, inpaint the background automatically, and take a guess at the back of the subject?
![F.HERO Ft. JSPKK x ลำไย ไหทองคำ x M-PEE - ไม่สนิทบิดหมด (Thai Riders Anthem) [Official MV]](http://i.ytimg.com/vi/kPJZ0UihBfk/mqdefault.jpg)
Hey bro! Impressing work! Is there any way I can consult with you on how to use that depth map for my project?
What i dont understand is that by now why dont we have the ability to have it guess at at least some of the backs of the objects, or to separate the subject from the background, inpaint the background automatically, and take a guess at the back of the subject?
There's are some attempts at this, MVDream and zero123 are both very interesting. But some of their examples seem cherrypicked.
Came for the flower, stayed for the algorithm.
How did you edit noise for Marigold?
It's a little bit of a mess right now, but I've added an example here github.com/graemeniedermayer/Marigold-Inpainting/tree/main
@@grae_n thanks so much. I go through that code. I love your channel and enthusiasm. Watched every video you post.