Omni6DPose / GenPose2

[ECCV 2024] GenPose++: A generative category-level 6D object pose estimation and tracking approach proposed in Omni6DPose.
https://jiyao06.github.io/Omni6DPose/
MIT License
23 stars 1 forks source link

RGBD data #5

Open tungyen opened 1 month ago

tungyen commented 1 month ago

If the input is RGB image and the partial point cloud from the depth image, is this input acceptable for GenPose?

Jiyao06 commented 1 month ago

Sure, GenPose takes as input the detection and segmentation results of a single object, along with the corresponding partial point cloud.

tungyen commented 1 month ago

Thank you for the reply. And I also want to ask if GenPose++ is able to operate on unknown data.

Jiyao06 commented 1 month ago

Yes, GenPose++ can generalize to unknown objects, and we will release a convenient version of GenPose++ with SAM for the downstream tasks.