AI Paper Atlas

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

DANet: Divergent Activation for Weakly Supervised Object Localization

Haolan Xue, Chang Liu, Fang Wan, Jianbin Jiao, Xiangyang Ji, Qixiang Ye

ICCV 2019 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised object localization · 225 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 hierarchical divergent activation (HDA) to CAM (ablation)VGG16 CAM · CUB-200-2011top-1 loc. err. (%): 55.85 -> 50.71Improved
add HDA to CAM (ablation)VGG16 CAM · CUB-200-2011top-1 cls. err. (%): 23.42 -> 24.13Hurt
add discrepant divergent activation (DDA) on top of HDA (ablation)VGG16 HDA · CUB-200-2011top-1 loc. err. (%): 50.71 -> 47.48Improved
add DDA on top of HDA (ablation)VGG16 HDA · CUB-200-2011top-5 loc. err. (%): 43.48 -> 38.04Improved
add multi-scale losses on a feature pyramid without DA modules (ablation)VGG16 CAM · CUB-200-2011top-1 loc. err. (%): 55.85 -> 58.58Hurt
add more hierarchy levels of supervision (one, two, three) (ablation)HDA · CUB-200-2011loc. err. (%): 55.85 -> 52.80 -> 50.71Improved
randomly drop half of the discrepant activation maps in each mini-batch (ablation)DDA · CUB-200-2011localizationImproved
set K=8 activation maps and lambda=0.01 (ablation)DDA · CUB-200-2011loc. err.Improved
use DANet instead of ACoL on VGG16VGG16 · CUB-200-2011top-1 loc. err. (%): 54.08 -> 47.48Improved
use DANet instead of SPG on VGG16VGG16 · CUB-200-2011top-1 loc. err. (%): 51.07 -> 47.48Improved
use DANet instead of ACoL on GoogLeNetGoogLeNet · ILSVRC validationtop-1 loc. err. (%): 53.28 -> 52.47Improved
use DANet instead of SPG on GoogLeNetGoogLeNet · ILSVRC validationtop-1 loc. err. (%): 51.40 -> 52.47No clear effect

Later papers that compared with this method

31 later accepted papers report DANet as a baseline in their results tables; 28 of them report a better number on most of the results they share with it.

23 more not listed.

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