From fbaaadd01e315c257d47d161e3a091475bdc8c57 Mon Sep 17 00:00:00 2001 From: Claire Chen Date: Tue, 2 Apr 2024 11:34:36 -0700 Subject: [PATCH] Update README.md --- aograsp_dataset_utils/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/aograsp_dataset_utils/README.md b/aograsp_dataset_utils/README.md index 3159657..5c1854d 100644 --- a/aograsp_dataset_utils/README.md +++ b/aograsp_dataset_utils/README.md @@ -1,6 +1,6 @@ # AO-Grasp dataset -The AO-Grasp dataset contains 78,000 6 DoF parallel jaw grasps on 84 articulated object instances across 7 categires from the PartNet-Mobility dataset. It contains grasps for each object in 10 joint states: 1 closed state and 9 randomly-sampled open states. For each object state, we provide the full point clouds and partial point clouds captured from 20 randomly-sampled camera viewpoints, as well as part segmentation masks. Additionally, we include the pre-processed, PartNet-Mobility objects we used to generated data. We have pre-processed these instances by running V-HACD on their meshes to obtain convex meshes, which we find result in better collision geometries in PyBullet. +The AO-Grasp dataset contains 78,000 6 DoF parallel jaw grasps on 84 articulated object instances across 7 categories from the PartNet-Mobility dataset. It contains grasps for each object in 10 joint states: 1 closed state and 9 randomly-sampled open states. For each object state, we provide the full point clouds and partial point clouds captured from 20 randomly-sampled camera viewpoints, as well as part segmentation masks. Additionally, we include the pre-processed, PartNet-Mobility objects we used to generated data. We have pre-processed these instances by running V-HACD on their meshes to obtain convex meshes, which we find result in better collision geometries in PyBullet. ## Downloading the AO-Grasp dataset