AI Paper Atlas

Research map / Detection and segmentation / CNN features

Densely Supervised Grasp Detector (DSGD)

Umar Asif, Jianbin Tang, Stefan Harrer

AAAI 2019 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 56 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
combine global, pixel-level, and region-level grasp predictions with confidence-based selectionDSGD · Cornell grasp datasetaccuracy (%): 96.8 -> 97.5Improved
combine hierarchical grasp predictionsDSGD · Multi-object datasetgrasp accuracy: 94.7% -> 97.2%Improved
use DenseNet rather than ResNet as the base networkDSGD · Cornell grasp datasetaccuracy (%): 96.7 -> 97.5Improved
use the lite network configuration (ablation)DSGD · Cornell grasp datasetaccuracy (%): 97.5 -> 97.1Hurt
use the lite network configuration (ablation)DSGD · Cornell grasp datasetspeed (fps): 9 -> 12Improved
increase dense-block widths using DSGD-A configuration (ablation)DSGD · Cornell grasp datasetaccuracy (%): 97.1 -> 97.1No clear effect
increase dense-block widths using DSGD-B configuration (ablation)DSGD · Cornell grasp datasetaccuracy (%): 97.1 -> 97.3Improved
increase growth rate to 48 in GGN and PGN (ablation)DSGD · Cornell grasp datasetaccuracy (%): 97.1 -> 97.5Improved
increase growth rate to 48 in GGN and PGN (ablation)DSGD · Cornell grasp datasetspeed (fps): 12 -> 9Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy