search

houph4 / Efficient6D-SAM-VLM-to-Grasping-task

Our repo containes a Efficient RGB-D features extractor to category-level and instance-level 6D pose estimation.
MIT License
3 stars 0 forks source link

readme

Efficient6D-SAM-VLM-to-Grasp

Our Repo is constantly being updated

Repo list :

  1. An Efficient RGB-D features extractor to category-level and instance-level 6D pose estimation.
  2. Bidirectional cross attentin fusion with rgb-point cloud and globe-local features.
  3. An end-to-end category level pose estimation inference scripts that includes real-time segmentation of SAM. (Free of shape piror)
  4. SAM-VLM to grasping task with 3DBBOX (Using Azure Kinect or Realsense).

Figure_1 Figure_2 Figure_3 output_video