AI Paper Atlas

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

A Weakly Supervised Amodal Segmenter With Boundary Uncertainty Estimation

Khoi Nguyen, Sinisa Todorovic

ICCV 2021 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 34 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
add occlusion boundary input instead of occluder mask (ablation)PCNet-m · COCOA-valmIoU: 80.73 -> 82.85Improved
add occlusion boundary input instead of occluder mask (ablation)PCNet-m · COCOA-valO-Acc: 85.75 -> 89.01Improved
add occlusion boundary input instead of occluder mask (ablation)PCNet-m · KINS-testinv-mIoU: 59.24 -> 62.77Improved
add uncertainty-weighted loss (ablation)PCNet-m · COCOA-valmIoU: 80.73 -> 82.49Improved
add uncertainty-weighted loss (ablation)PCNet-m · COCOA-testO-Acc: 86.73 -> 88.40Improved
add uncertainty-weighted loss (ablation)PCNet-m · KINS-testinv-mIoU: 59.24 -> 62.00Improved
use proposed uncertainty data term rather than uncertainty-weighted BCE (ablation)ASBU · COCOA-valmIoU: 83.18 -> 84.22Improved
use proposed uncertainty data term rather than uncertainty-weighted BCE (ablation)ASBU · KINS-testmIoU: 94.34 -> 94.83Improved
combine boundary input and uncertainty loss (full ASBU)PCNet-m · COCOA-testmIoU: 86.63 -> 89.87Improved
combine boundary input and uncertainty loss (full ASBU)PCNet-m · KINS-testmIoU: 94.52 -> 94.83No clear effect
train Mask-RCNN on ASBU pseudo masks instead of PCNet-m pseudo masksMask-RCNN · COCOA-testAP: 22.6 -> 23.8Improved
train Mask-RCNN on ASBU pseudo masks instead of PCNet-m pseudo masksMask-RCNN · KINS-testAP: 29.1 -> 29.3No clear effect

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