Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

Visual Grasp Affordances From Appearance-Based Cues

Hyun Oh Song, Mario Fritz, Chunhui Gu and Trevor Darrell

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2013-16
March 4, 2013

http://www.eecs.berkeley.edu/Pubs/TechRpts/2013/EECS-2013-16.pdf

In this paper, we investigate the prediction of visual grasp affordances from 2D measurements. Appearance- based estimation of grasp affordances is desirable when 3- D scans are unreliable due to clutter or material proper- ties. We develop a general framework for estimating grasp affordances from 2-D sources, including local texture-like measures as well as object-category measures that capture previously learned grasp strategies. Local approaches to estimating grasp positions have been shown to be effective in real-world scenarios, but are unable to impart object- level biases and can be prone to false positives. We de- scribe how global cues can be used to compute continu- ous pose estimates and corresponding grasp point loca- tions, using a max-margin optimization for category-level continuous pose regression. We provide a novel dataset to evaluate visual grasp affordance estimation; on this dataset we show that a fused method outperforms either local or global methods alone, and that continuous pose estimation improves over discrete output models.

Advisor: Trevor Darrell


BibTeX citation:

@mastersthesis{Song:EECS-2013-16,
    Author = {Song, Hyun Oh and Fritz, Mario and Gu, Chunhui and Darrell, Trevor},
    Title = {Visual Grasp Affordances From Appearance-Based Cues},
    School = {EECS Department, University of California, Berkeley},
    Year = {2013},
    Month = {Mar},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/2013/EECS-2013-16.html},
    Number = {UCB/EECS-2013-16},
    Abstract = {In this paper, we investigate the prediction of visual grasp affordances from 2D measurements. Appearance- based estimation of grasp affordances is desirable when 3- D scans are unreliable due to clutter or material proper- ties. We develop a general framework for estimating grasp affordances from 2-D sources, including local texture-like measures as well as object-category measures that capture previously learned grasp strategies. Local approaches to estimating grasp positions have been shown to be effective in real-world scenarios, but are unable to impart object- level biases and can be prone to false positives. We de- scribe how global cues can be used to compute continu- ous pose estimates and corresponding grasp point loca- tions, using a max-margin optimization for category-level continuous pose regression. We provide a novel dataset to evaluate visual grasp affordance estimation; on this dataset we show that a fused method outperforms either local or global methods alone, and that continuous pose estimation improves over discrete output models.}
}

EndNote citation:

%0 Thesis
%A Song, Hyun Oh
%A Fritz, Mario
%A Gu, Chunhui
%A Darrell, Trevor
%T Visual Grasp Affordances From Appearance-Based Cues
%I EECS Department, University of California, Berkeley
%D 2013
%8 March 4
%@ UCB/EECS-2013-16
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/2013/EECS-2013-16.html
%F Song:EECS-2013-16