AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Pull the Plug? Predicting If Computers or Humans Should Segment Images

Danna Gurari, Suyog Jain, Margrit Betke, Kristen Grauman

CVPR 2016 · Detection and segmentation › Open-set and few-shot detection · approach: Amodal and open-world instance segmentation · 24 citations (OpenAlex)

Check 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
Predict segmentation quality from binary-mask features instead of CPMC featuresSegmentation-quality prediction · BU-BILCC: 0.36 -> 0.61Improved
Predict segmentation quality from binary-mask features instead of CPMC featuresSegmentation-quality prediction · WeizmannMAE: 0.32 -> 0.24Improved
Use AlexNet CNN features for prediction instead of binary-mask featuresSegmentation-quality prediction · BU-BIL, Weizmann, and IISCC: 0.61 -> -0.01; 0.64 -> -0.1; 0.68 -> -0.01Hurt
Train and test on the same dataset rather than cross-set trainingSegmentation-quality prediction · BU-BILCC: 0.61 -> 0.69Improved
Train and test on the same dataset rather than cross-set trainingSegmentation-quality prediction · IISCC: 0.68 -> 0.78Improved
Select computer initializations using predicted quality and allocate humans to the lowest-scoring imagesPTP coarse-segmentation initialization · Weizmann, IIS, and BU-BIL; segmentation-tool initializationhuman annotation effort eliminated: 44%Improved
Use predicted-quality allocation rather than the Rectangle baselineLankton level sets initialization · Weizmann, IIS, and BU-BILsegmentation quality: 20 percentage pointsImproved
Use predicted-quality allocation rather than relying exclusively on human initializationPTP coarse-segmentation initialization · Weizmann, IIS, and BU-BILsegmentation quality: six percentage point improvementImproved
Allocate humans to low-predicted-quality computer segmentations and retain computer results elsewherePTP fine-grained segmentation · Weizmann, IIS, and BU-BILhuman annotation time saved at comparable segmentation quality: 30-60 minutes less human annotation timeImproved
Use computer segmentations instead of crowd annotations when predictions favor the computerPTP fine-grained segmentation · Last 100 images assigned human annotationsimages where computer segmentation has higher Jaccard than crowd: 10%Improved

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