search

graspnet / graspnet-baseline

Baseline model for "GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping" (CVPR 2020)
https://graspnet.net/
Other
415 stars 133 forks source link

Questions about the scene 6DOF pose annotation. #75

Closed qiaoyo closed 9 months ago

qiaoyo commented 10 months ago

Hi, Thanks for your great work and detailed API tool. I have received the Grasp pose annotation tool invitation. I noticed that it's basically related to PointnetGPD and integrated into the GraspNetAPI. Besides the object pose annotation, the 6DOF pose annotation is also an important part in your great work. In your paper, this part is described in this way "we only need to annotate 6D poses for the first frame of each scene." I want to know how this part is achieved. Some tools I have known is the Elasticfusion and Labelfusion. The former is to rebuild the scene pointcloud, the latter is to annotate the 6DOF pose. Could u share the 6DOF annotation way of the first frame? Sincerely, thanks very much.

chenxi-wang commented 9 months ago

Hi, we made a tool to render the objects to the image, in which way we can visualize and adjust 6dof poses easily.

willieChen58 commented 8 months ago

Hi, we made a tool to render the objects to the image, in which way we can visualize and adjust 6dof poses easily.

@chenxi-wang Hi, could you provide with the tools you made to annotate 6D poses? Because our scientific research project needs to use 6D pose estimation and grasp, we need to label the dataset of pose estimation, and there is no mature method of annotating the dataset, so I would like to ask if you can provide this annotation tool?