This is the dataset part for our accepted ICRA 2020 paper:
PointNet++ Grasping: Learning An End-to-end Spatial Grasp Generation Algorithm from Sparse Point Clouds
Video：https://www.youtube.com/watch?v=AfU7npscnZ0
Dataset:
Link：https://pan.baidu.com/s/1_prfq4A_Dg9kREqpc3Cikw Password：j7zn
The object models can be downloaded from http://ycb-benchmarks.s3-website-us-east-1.amazonaws.com/
The code is to visualize the performance of Single-object Grasp Planning. We refer dexnet(https://berkeleyautomation.github.io/dex-net/)
The installation method is the same with dexnet (https://berkeleyautomation.github.io/dex-net/install/install.html)
You can run show show_labels.py to get the following cases:
(The blue grasp denotes the major grasp and the rest are supplementary grasps)
