AI Paper Atlas

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

Generative Adversarial Learning Towards Fast Weakly Supervised Detection

Yunhan Shen, Rongrong Ji, Shengchuan Zhang, Wangmeng Zuo, Yan Wang

CVPR 2018 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised detection · 83 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 both adversarial and structural similarity losses instead of adversarial loss alone (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 39.2 -> 44.3Improved
Add 128 feedback proposals (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 44.3 -> 45.8Improved
Add 256 feedback proposals (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 44.3 -> 46.5Improved
Add 512 feedback proposals (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 44.3 -> 47.0Improved
Increase feedback proposals from 512 to 1024 (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 47.0 -> 47.1Improved
Use feedback proposals rather than original proposals for the additional proposal set (ablation)GAL-fWSD300 · PASCAL VOC 2007 testmAP: 44.3 -> 47.0Improved
Use 512 × 512 input instead of 300 × 300 input (ablation)GAL-fWSD · PASCAL VOC 2007 testmAP: 47.0 -> 47.5Improved
Use 512 × 512 input instead of 300 × 300 input (ablation)GAL-fWSD · PASCAL VOC 2007 trainCorLoc: 68.1 -> 66.1Hurt

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