The Columbia grasp database

By Corey Goldfeder, Matei T. Ciocarlie, Hao Nguyen Dang, Peter K. Allen. 2009

353 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Control and Systems Engineering, Evaluation & Benchmarks, Human-Computer Interaction, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Control and Systems Engineering

In Evaluation & Benchmarks

In Human-Computer Interaction

In Robotics & Control

Citations in the corpus

Learning Synergies Between Pushing and Grasping with Self-Supervised Deep Reinforcement Learning (2018, background): “Grasping too, has been well studied in the domain of model-based reasoning [19]; from modeling contact forces and their resistance to external wrenches [20], [21], to characterizing grasps by their ability to constrain object mobility [22].”
Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection (2016, methodology): “These methods typically need to understand the geometry of the scene, using depth or stereo sensors and matching of previously scanned models to observations (Goldfeder et al., 2009b).”
Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours (2015, methodology): “For gripping we use the stock two ﬁngered parallel gripper with a maximum width (open state) of 75 mm and a minimum width (close state) of 37 mm.”
IRIS: Implicit Reinforcement without Interaction at Scale for Learning Control from Offline Robot Manipulation Data (2019, background): “large amounts of data for tasks such as grasping in both simulated [13, 18, 25] and physical settings [17, 22, 29].”
Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning (2019, background): “Similarly, there exists a number of robotics simulation benchmarks including visual navigation [39, 40, 41, 42, 43], autonomous driving [44, 45, 46], grasping [47, 48, 49], single-task manipulation [50], among others.”
ROBOTURK: A Crowdsourcing Platform for Robotic Skill Learning through Imitation (2018, methodology): “SSL has been used successfully to collect and learn from large quantities of data in both simulated [7, 8, 9] and physical settings [5, 10, 11] for tasks such as grasping.”
Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics (2017, methodology): “To execute grasps on a physical robot, a common approach is to precompute a database of known 3D objects labeled with grasps and quality metrics such as GraspIt!”
Data-Driven Grasp Synthesis—A Survey (2013, methodology): “Although various object-grasp databases are already available e.g. the Columbia Grasp database [140], the VisGraB data set [141] or the playpen data set [142] they are not commonly used for comparison.”
Grasping in the Wild: Learning 6DoF Closed-Loop Grasping From Low-Cost Demonstrations (2019, background): “Classic vision-based grasping solutions often explicitly model contact forces with prior knowledge of object geometry, pose, and dynamics [25, 31, 6, 38].”
Benchmarking in Manipulation Research: Using the Yale-CMU-Berkeley Object and Model Set (2015, methodology): “The Columbia grasp database [3] rearranges the object models of the Princeton shape benchmark [18] for robotic manipulation and provides mesh models of 8,000 objects together with a number of successful grasps per model.”
Dex-Net 1.0: A cloud-based network of 3D objects for robust grasp planning using a Multi-Armed Bandit model with correlated rewards (2016, methodology): “[12], [13] developed the Columbia grasp database, a dataset of 1,814 distinct models and over 200,000 force closure grasps generated using the GraspIt! sampling-based grasp planner.”
Learning Visual Affordances for Robotic Manipulation (2019, methodology): “A common approach to deploying these methods in real systems involves pre-computing grasps from a database of known 3D object models [34], and indexing them at run-time with point cloud registration for object pose estimation [123, 122].”
Pose error robust grasping from contact wrench space metrics (2012, methodology): “The grasps we have analyzed in this work are drawn from the Columbia Grasp Database.”
Autonomously Acquiring Instance-Based Object Models from Experience (2015, background): “Modern approaches use object recognition systems to estimate pose and object type, then libraries of grasps either annotated or learned from data [36, 8, 30].”
The OpenGRASP benchmarking suite: An environment for the comparative analysis of grasping and dexterous manipulation (2011, background): “…and provides a related categorization of objects [19] and web-based data bases which provide object models for grasping along with sets of stable grasps, namely the Columbia Grasp Database [20], the KIT ObjectModels Web Database [21], [22] and the selection of every day objects described in [23].”
Learning to grasp under uncertainty (2011, background): “And it also verifies that T1 and H1 are “cases where a grasp satisfies our quality metrics, but would require a degree of precision that cannot be obtained in real-life execution” [ 7 ].”
Efficient grasping from RGBD images: Learning using a new rectangle representation (2011, background): “Another notable work is GraspIt [19, 20] where a large database of 3D objects was used to evaluate the grasps.”
Learning Joint Reconstruction of Hands and Manipulated Objects (2019, positive): “Following [9], our generated grasps optimize for the grasp metric but do not necessarily reflect the statistical distribution of human grasps.”
Learning Grasping Interaction with Geometry-aware 3D Representations (2017, background): “Some work focused on analytic modeling of robotic grasps with known object shape information Goldfeder et al. [2009], Leon et al. [2010].”