AI Paper Atlas

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

Regularized Loss for Weakly Supervised Single Class Semantic Segmentation

Olga Veksler

ECCV 2020 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 15 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
start from OxfordPet-pretrained weights instead of annealing (ablation)UMobV2 regularized loss · ECSSDF-measure: 57.96 -> 79.20Improved
start from OxfordPet-pretrained weights instead of annealing (ablation)UResNext regularized loss · ECSSDF-measure: 73.47 -> 83.49Improved
start from OxfordPet-pretrained weights instead of annealing (ablation)UResNext regularized loss · DUT-OMRONF-measure: 64.48 -> 73.42Improved
start from OxfordPet-pretrained weights instead of annealing (ablation)UMobV2 regularized loss · MSRA-BF-measure: 86.20 -> 85.51Hurt
start from OxfordPet-pretrained weights instead of annealing (ablation)UMobV2 regularized loss · DUT-OMRONF-measure: 53.79 -> 54.49No clear effect
add annealing of sigma and complete loss (all parts) (ablation)regularized loss training · single-class datasetstraining successImproved
use larger UResNext backbone instead of UMobV2regularized loss · DUT-OMRONF-measure: 53.79 -> 64.48Improved
train with ground truth instead of regularized lossUMobV2 · ECSSDF-measure: 57.96 -> 84.78Improved
add negative loss using other categories as negatives (ablation)UnetX · Pascal VOC cosegmentationJaccard (mean): 0.59 -> 0.62Improved
train on larger Web dataset instead of the small cosegmentation setUnetX with negative loss · Pascal VOC cosegmentationJaccard (mean): 0.62 -> 0.67Improved
add negative loss, train on PascalSingle (ablation)DeepLab-ResNet101 pseudo-labels · Pascal VOC 2012 valmIoU: 66.7 -> 67.1Improved
train on PascalSingle instead of Web imagesDeepLab-ResNet101 pseudo-labels · Pascal VOC 2012 valmIoU: 64.0 -> 66.7Improved

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