AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

Block Annotation: Better Image Annotation With Sub-Image Decomposition

Hubert Lin, Paul Upchurch, Kavita Bala

ICCV 2019 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 22 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
annotate 50% of blocks per image instead of full imagesfull annotation · Cityscapes val (all images)mIoU: 77.7 -> 77.7No clear effect
annotate 50% of blocks per image instead of full imagesfull annotation · ADE20K val (all images)mIoU: 37.4 -> 37.2No clear effect
annotate 12% of blocks per image instead of full imagesfull annotation · Cityscapes val (all images)mIoU: 77.7 -> 74.6Hurt
annotate 12% of blocks per image instead of full imagesfull annotation · ADE20K val (all images)mIoU: 37.4 -> 36.1Hurt
use block annotation instead of full-image annotation at equal pixel budgetfull annotation · Cityscapes and ADE20Kerror reduction: 13% (10%) error reduction at 12% budgetImproved
use pseudo-checkerboard block selection instead of checkerboard (ablation)checkerboard blocks · Cityscapes and ADE20KmIoUImproved
use block annotation instead of scribbles at equal timeScribbleSup · Pascal VOC 2012 valmIoU: 63.1 -> 67.2Improved
use block annotation instead of coarse annotation at equal timeCoarse annotation · Cityscapes valmIoU: 68.8 -> 72.1Improved
use block annotation instead of full-image annotation for crowdsourcingFull-image annotation · SUNCG annotation errorerror rate: 0.286 -> 0.253Improved
use block annotation instead of full-image annotation for crowdsourcingFull-image annotation · SUNCG annotation error (small regions)error rate: 0.677 -> 0.636Improved
block inpainting instead of automatic segmentationAutomatic segmentation · Cityscapes val (Block-50%)mIoU: 78 -> 92Improved
block inpainting instead of automatic segmentationAutomatic segmentation · Cityscapes val (Block-12%)mIoU: 75 -> 82Improved

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