AI Paper Atlas

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

Incorporating Network Built-in Priors in Weakly-Supervised Semantic Segmentation

Fatemeh Sadat Saleh, Mohammad Sadegh Aliakbarian, Mathieu Salzmann, Lars Petersson, José Manuel Álvarez, Stephen Jay Gould

TPAMI 2017 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 41 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 foreground/background masks from pretrained VGG-16 activations (ablation)tag-only segmentation network · PASCAL VOC 2012 valmIoU: 31.0 -> 47.3Improved
add multi-class masks (localization + fusion) instead of no mask (ablation)tag-only segmentation network · PASCAL VOC 2012 valmIoU: 31.0 -> 49.2Improved
add higher-order CRF term to fg/bg masks (ablation)fg/bg mask method · PASCAL VOC 2012 valmIoU: 47.3 -> 48.0Improved
add higher-order CRF term to localization masks (ablation)localization prior method · PASCAL VOC 2012 valmIoU: 45.9 -> 46.6Improved
add fusion with fg/bg prior to localization masks (ablation)localization prior method · PASCAL VOC 2012 valmIoU: 45.9 -> 49.2Improved
use large field of view instead of small (ablation)multi-class mask method with higher-order term · PASCAL VOC 2012 valmIoU: 49.3 -> 50.9Improved
use built-in fg/bg masks instead of external objectness masksfg/bg mask generation · PASCAL VOC 2012 valmean IoU of masks: 52.34% -> 60.08%Improved
add fusion to localization masks used directly without training (ablation)multi-class masks · PASCAL VOC 2012 valmIoU: 43.0 -> 46.6Improved
train a network on masks instead of predicting from masks directlymulti-class masks · PASCAL VOC 2012 valmIoU: 46.6 -> 50.9Improved
use multi-class masks instead of fg/bg masksOurs · PASCAL VOC 2012 testmIoU: 49.6 -> 52.6Improved
use our method instead of SECweakly supervised segmentation · MS COCO valmIoU: 22.4 -> 20.4Hurt
use our method instead of SECweakly supervised segmentation · YTOmIoU: 48.6 -> 51.6Improved

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