AI Paper Atlas

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

Hide-And-Seek: Forcing a Network to Be Meticulous for Weakly-Supervised Object and Action Localization

Krishna Kumar Singh, Yong Jae Lee

ICCV 2017 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised object localization · 737 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
hide random 32x32 patches during trainingAlexNet-GAP · ILSVRC valGT-known Loc: 54.90 -> 58.75Improved
hide random 32x32 patches during trainingGoogLeNet-GAP · ILSVRC valTop-1 Loc: 43.60 -> 45.21Improved
hide patches during trainingAlexNet-GAP · ILSVRC valTop-1 Clas: 60.23 -> 57.94Hurt
use mixed patch sizes (16,32,44,56 or none) per image (ablation)AlexNet-HaS · ILSVRC valTop-1 Loc: 37.34 -> 37.65Improved
ensemble HaS models of several patch sizes instead of five GAP models (ablation)AlexNet · ILSVRC valGT-known Loc: 56.91 -> 60.14Improved
hide patches instead of image-level 50% dropout (ablation)AlexNet · ILSVRC valTop-1 Loc: 31.68 -> 37.65Improved
use global max pooling with Hide-and-Seek instead of GAP baseline (ablation)AlexNet · ILSVRC valGT-known Loc: 50.40 -> 59.27Improved
max pooling instead of average pooling under Hide-and-Seek (ablation)AlexNet-HaS-56 · ILSVRC valGT-known Loc: 58.43 -> 59.27Improved
hide patches in conv1 feature map (size 11) (ablation)AlexNet-GAP · ILSVRC valGT-known Loc: 54.90 -> 58.33Improved
raise hiding probability from 50% to 75% (ablation)AlexNet-HaS-56 · ILSVRC valTop-1 Loc: 37.34 -> 34.21Hurt
lower hiding probability from 50% to 25% (ablation)AlexNet-HaS-56 · ILSVRC valGT-known Loc: 58.43 -> 57.49Hurt
hide frame segments during trainingC3D action classifier · THUMOS 2014 val (IoU 0.5)mAP: 6.11 -> 6.84Improved

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