AI Paper Atlas

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

Keep CALM and Improve Visual Feature Attribution

Jae Myung Kim, Junsuk Choe, Zeynep Akata, Seong Joon Oh

ICCV 2021 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised object localization · 20 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
replace CAM with CALM_EM attributionResNet50 · CUB class pairs attribution (mean)mPxAP (%): 29.6 -> 34.3Improved
train with EM instead of marginal likelihood (ablation)CALM · CUB class pairs attribution (mean)mPxAP (%): 27.3 -> 34.3Improved
replace CAM with CALM_ML attributionResNet50 · CUB class pairs attribution (mean)mPxAP (%): 29.6 -> 27.3Hurt
replace CAM-style classifier with CALM_EMResNet50 · ImageNet classificationaccuracy (%): 74.5 -> 70.4Hurt
replace CAM-style classifier with CALM_MLResNet50 · CUB classificationaccuracy (%): 70.6 -> 59.6Hurt
replace CAM-style classifier with CALM_EMResNet50 · CUB classificationaccuracy (%): 70.6 -> 71.8No clear effect
add class-hierarchy superset aggregation (+Y) (ablation)CALM_EM · CUB WSOLMaxBoxAccV2 (%): 52.5 -> 65.4Improved
add class-hierarchy superset aggregation (+Y) (ablation)CALM_ML · CUB WSOLMaxBoxAccV2 (%): 61.3 -> 68.0Improved
add class-hierarchy superset aggregation (+Y) (ablation)CAM · CUB WSOLMaxBoxAccV2 (%): 61.1 -> 63.4Improved
add class-hierarchy superset aggregation (+Y) (ablation)CALM_EM · ImageNet WSOLMaxBoxAccV2 (%): 62.5 -> 62.8No clear effect
add class-hierarchy superset aggregation (+Y) (ablation)CAM · ImageNet WSOLMaxBoxAccV2 (%): 62.4 -> 60.6Hurt
add class-hierarchy superset aggregation (+Y) (ablation)CALM_EM · OpenImages WSOLPxAP (%): 62.7 -> 62.7No 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