Yuzhe Qin

> Thank you very much @jackjansen!!! Your comment helped me solve my issue with the build of my Dockerfile. > > Looks like there are newer .debs available: > >...

hi @candace3698 It is the transform from world coordinate to the grasp coordinate. The xyz axis definition follow the same convention as previous work like [GPD](https://arxiv.org/abs/1706.09911) and [PointNetGPD](https://arxiv.org/abs/1809.06267) ee, ee...

The network work files are included in the repo. The entry code for training will be public soon.

Hi @maximiliangilles and @H-HChen, Thanks for your question and interest. I am sorry that the project and code is developed in 2019 and it has been a long time since...

https://drive.google.com/file/d/184tnRAHz9dnzhz86Jfcs35t7mCr8BlE5/view?usp=sharing

Hello @taewookim0812, `hand_frame_seq` refers to the sequence of hand poses derived from detectors that utilize the MANO framework. Since MediaPipe operates independently of the MANO/SMPL-X model, it doesn't provide such...

Hi @Charlie0257 , The point cloud video showcased on the project page is created using the [Mitsuba Renderer](https://github.com/tolgabirdal/Mitsuba2PointCloudRenderer) to display the point cloud. I'm confident that a comparable result is...

Hi @wslgqq277g There is no pre-trained PointNet in this project. However, the PointNet architecture of DexPoint is the same as the one used in DexArt. The PointNet here is trained...