AI Paper Atlas

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

Phase Concentration and Shortcut Suppression for Weakly Supervised Semantic Segmentation

Hoyong Kwon, Jaeseok Jeong, Sung-Hoon Yoon, Kuk-Jin Yoon

ECCV 2024 · Detection and segmentation › Weakly supervised detection and segmentation · approach: CAM refinement and prototypes · 5 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 Frequency Shortcut Suppression (FSS) (ablation)ViT WSSS baseline (refined CAMs) · PASCAL VOC 2012 trainmIoU (%): 63.2 -> 65.4Improved
add Magnitude-mixing-based Phase Concentration (MPC) (ablation)ViT WSSS baseline (refined CAMs) · PASCAL VOC 2012 trainmIoU (%): 63.2 -> 67.2Improved
add FSS on top of MPC (ablation)baseline + MPC · PASCAL VOC 2012 trainmIoU (%): 67.2 -> 69.5Improved
add FSS to raw (unrefined) CAMs (ablation)baseline · PASCAL VOC 2012 trainprecision / mIoU: precision 53.2 -> 71.3; mIoU 45.3 -> 59.4Improved
add FSS (ablation)bird class CAMs · PASCAL VOC 2012 trainprecision: 17.3 -> 59.2Improved
apply MPC at layer 8 instead of late layers 10 or 11 (ablation)MPC · PASCAL VOC 2012 trainmIoU (%): 66.4 (layer 11) -> 69.5 (layer 8)Improved
apply MPC at early layers (3, 4) instead of 8 (ablation)MPC · PASCAL VOC 2012 trainmIoU (%): 69.5 -> 68.0 (layer 3)Hurt
use magnitude mixing ratio r below 0.5 (best 0.3) (ablation)MPC · PASCAL VOC 2012 trainmIoU (%): minimum gain of 2.3%p over FSS aloneImproved
use ours with DeiT-S pseudo labels instead of USAGEWSSS segmentation · PASCAL VOC 2012 valmIoU (%): 71.9 -> 73.2Improved
use ours instead of ACRWSSS segmentation · MS COCO 2014 valmIoU (%): 45.3 -> 45.7Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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