AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Semantic Understanding of Scenes Through the ADE20K Dataset

Bolei Zhou, Hang Zhao, Xavier Puig, Tete Xiao, Sanja Fidler, Adela Barriuso, Antonio B. Torralba

IJCV 2018 · Detection and segmentation › Open-set and few-shot detection · approach: Amodal and open-world instance segmentation · 2,004 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
use synchronized BN with batch/BN size 16 instead of frozen BN (ablation)PSPNet DilatedResNet-50 · SceneParse150 valmean IoU: 3.17%Improved
reduce BN size from 4 to 2 (ablation)PSPNet DilatedResNet-50 · SceneParse150 valmean IoU: 0.4119 -> 0.3355Hurt
change batch size from 16 to 4 with synchronized BN (ablation)PSPNet DilatedResNet-50 · SceneParse150 valmean IoU: 0.4126 -> 0.4119No clear effect
increase batch size to 16 without synchronizing BN (BN size 2) (ablation)PSPNet DilatedResNet-50 · SceneParse150 valmean IoU: 0.3355 -> 0.3403No clear effect
vary batch size with frozen BN (ablation)PSPNet DilatedResNet-50 · SceneParse150 valmean IoU: 0.3809 -> 0.3856No clear effect
add cascade framework to SegNetSegNet · SceneParse150 valmean IoU: 0.2164 -> 0.2751Improved
add cascade framework to DilatedVGGDilatedVGG · SceneParse150 valmean IoU: 0.3231 -> 0.3490Improved
add Pyramid Pooling Module to DilatedResNet-50 (ablation)DilatedResNet-50 · SceneParse150 valmean IoU: 0.3549 -> 0.4204Improved
add Pyramid Pooling Module to DilatedResNet-18 (ablation)DilatedResNet-18 · SceneParse150 valmean IoU: 0.3534 -> 0.3800Improved
use multi-scale training for Mask R-CNN (ablation)Mask R-CNN FPN-50 · InstSeg100 valmAP: .1832 -> .2241Improved
fuse Mask R-CNN instance masks into scene parsing outputDilatedResNet-50 + PPM · SceneParse150 valmean IoU: 0.4204 -> 0.4256Improved
fuse Mask R-CNN instance masks into scene parsing outputDilatedResNet-50 + PPM · SceneParse150 valpixel accuracy: 80.23% -> 80.21%No 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