Jiazhong Cen

Yeah, rebalancing the positive and negative terms of the loss function can help (but may introduce other problem like segmenting different objects into one). The scale parameter is introduced to...

Hi, thanks for your interest. However, the number of points (3D Gaussians) is determined by the 3D-GS densification process automatically. A fact is that even if you have a very...

Hi, I'm not sure about what dataset do you mean. If you are asking about the lerf dataset, they used the nerf studio to convert the camera poses collected by...

Hi, the evaluation script is incompatible with this GUI version, which has gone through code refactoring and abandoned many interface for evaluation. The evaluation mainly involves loading different datasets like...

Hi, the reference view is set to the first frame of the sorted views. However, the method is robust to reference view selection since the segmentation target is relatively simple.

No, the target views include the reference view generally. Since though the reference view has gt mask for reference, the segmentation cannot ensure the final result align with the initial...

Hi. You can check the two scripts for details. The extract_segment_everything_masks.py uses the segment everything func of SAM to extract masks for input images one by one. And the get_scale.py...

Hello, can you install the original rasterizer provided by 3D-GS?

> rasterize_points.cu(118): error: argument of type "__nv_bool" is incompatible with parameter of type "int *" I guess there may be some wrongly defined vars like the newly added 'mask'. You...

Hi, the point cloud is stored as an additional `mask' attribute in the 3D-GS model, which is essentially same as the precomputed_mask.pt in SAGA.