AI Paper Atlas

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

Detecting 11K Classes: Large Scale Object Detection Without Fine-Grained Bounding Boxes

Hao Yang, Hao Wu, Hao Chen

ICCV 2019 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised detection · 23 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
train jointly with fine-grained weakly-labeled data (naive two-stream)SNIPER coarse-grained detector · OpenImages V4 valmAP-CG: 45.7 -> 52.5Improved
add soft-attention proposal re-ranking (ablation)naive two-stream · OpenImages V4 valmAP-FG: 34.0 -> 49.2Improved
add soft-attention proposal re-ranking (ablation)naive two-stream · OpenImages V4 valmAP-CG: 52.5 -> 52.9No clear effect
add single-level coarse-grained memory (ablation)+Soft-Attention · OpenImages V4 valmAP-FG: 49.2 -> 49.6No clear effect
add single-level coarse-grained memory (ablation)+Soft-Attention · OpenImages V4 valmAP-CG: 52.9 -> 52.2Hurt
replace single-level memory with dual-level memory with foreground attention pooling (ablation)+CG-Memory · OpenImages V4 valmAP-FG: 49.6 -> 51.9Improved
replace single-level memory with dual-level memory with foreground attention pooling (ablation)+CG-Memory · OpenImages V4 valmAP-CG: 52.2 -> 53.5Improved
add fine-grained weak stream with shared backbone vs pure weakly-supervised fine-tuning (ablation)naive two-stream · OpenImages V4 valmAP-FG: 20.2 -> 34.0Improved
add fine-grained weakly labeled 3K classes to coarse detectorSNIPER coarse-grained detector · ImageNetmAP-CG: 54.0 -> 50.7Hurt
scale fine-grained weak classes from 3K to 11KOurs · ImageNetmAP-CG: 50.7 -> 49.1Hurt
use weak-label semi-supervised model instead of full-supervision for fine-grainedOurs-3K vs SNIPER-FG-Fully · ImageNetmAP-FG: 41.6 -> 35.1Hurt
use weak-label model instead of full-supervision at stricter IoUOurs vs SNIPER-FG-Fully · ImageNetmAP@0.5:0.95 FG: 22 -> 25Improved

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