AI Paper Atlas

Research map / Detection and segmentation / CNN features

Expectation-Maximization Attention Networks for Semantic Segmentation

Xia Li, Zhisheng Zhong, Jianlong Wu, Yibo Yang, Zhouchen Lin, Hong Liu

ICCV 2019 · Detection and segmentation › CNN features · approach: Context modeling for semantic segmentation · 749 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
update initial bases by moving average instead of back-propagation or no update (ablation)EMAU · PASCAL VOC val (R50)mIoU: surpasses others by at least 0.9 in mIoUImproved
update initial bases by back-propagation instead of keeping them fixed (ablation)EMAU · PASCAL VOC val (R50)mIoUHurt
apply L2Norm to bases instead of no normalization (ablation)EMAU · PASCAL VOC val (R50)mIoUImproved
apply LayerNorm to bases instead of no normalization (ablation)EMAU · PASCAL VOC val (R50)mIoUHurt
increase evaluation iterations T_eval (ablation)EMAU · PASCAL VOC val (R50)mIoUImproved
increase training iterations T_train beyond 3 (ablation)EMAU · PASCAL VOC val (R50)mIoUHurt
EM iterations (3) versus single iteration A2 block (ablation)EMANet · PASCAL VOC val (R50)mIoU: 77.41 vs 77.34No clear effect
EM iterations versus non-local module (ablation)EMANet · PASCAL VOC val (R50)mIoU: 77.78 vs 77.34No clear effect
widen EMANet input channels from 256 to 512EMANet · PASCAL VOC val SSmIoU: 79.73 -> 80.05Improved
replace DeeplabV3 with EMANet (512), single scaleResNet-101 · PASCAL VOC val SSmIoU: 78.51 -> 80.05Improved
replace DeeplabV3 with EMANet (512), MS+FlipResNet-101 · PASCAL VOC val MS+FlipmIoU: 79.77 -> 81.32Improved
replace DeeplabV3 by EMANet on testResNet-101 · PASCAL VOC testmIoU: 85.7 -> 87.7Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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