gaussian-grouping icon indicating copy to clipboard operation
gaussian-grouping copied to clipboard

Gaussian Grouping for open-world Anything reconstruction, segmentation and editing.

Results 52 gaussian-grouping issues
Sort by recently updated
recently updated
newest added

Hi Thank you for your amazing work, Is it possible that I can implement your own method without the need to train the entire model again because I already have...

Recently, I came across your paper, and I have a question regarding Gaussian Grouping. Specifically, I found the Recomposition method mentioned in your paper to be quite intriguing, and I...

Thanks for your great work!! As I understanding, the subfolder in `test_mask` folder means the object masks of some frame in `images` folder. But I don't which frame in `images`...

Hi,there. I am trying the inpaint function. By following your steps listed, I finally miss the folder inpaint_object_mask_255 which i assume is the 255 mask for the removed object, any...

hi, I have some questions about how to implement panoptic segmentation, after you got semantic labels from Maskdino, did you add another linear layer like instance ,to project indentity embedding...

https://github.com/lkeab/gaussian-grouping/blob/0ab60afed3385b717c985af1d30a20f7b0884c89/scene/gaussian_model.py#L322 I guess you want to optimize some parameters of nn.parameter and fix some parameters during the optimization process, but I found that it seems that this is not feasible...

Hi, when I implement this '' bash script/prepare_pseudo_label.sh spinerf 1'' It then give me this mistake: UserWarning: Overwriting tiny_vit_21m_512 in registry with deva.ext.MobileSAM.tiny_vit_sam.tiny_vit_21m_512. This is because the name being registered...

Hi! excellent work! I'm new to this field. About the parameters of gaussian class, i'd like to know the meaning of "_objects_dc", "_features_dc" and "_features_rest".And what is their role? ![image](https://github.com/user-attachments/assets/9db596bb-bea9-41ac-af4e-57f525dee388)

Is this approach extendable into 3D gaussian-level clustering of gaussian. I mean without identity-alpha bleding and convolutional classifier layer. For example, if I would like to create a map where...

Am I right that in id renders of the reconstructed scene black color corresponds to non-tracked objects?