AI Paper Atlas

Research map / Detection and segmentation / CNN features

Instances As Queries

Yuxin Fang, Shusheng Yang, Xinggang Wang, Yu Li, Chen Fang, Ying Shan, Bin Feng, Wenyu Liu

ICCV 2021 · Detection and segmentation › CNN features · approach: Boundary-aware instance segmentation · 390 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 DynConv mask module (parallel dynamic mask heads) to vanilla query-based mask head (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 37.9 -> 39.8Improved
add parallel supervision on all stages with DynConv (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 38.8 -> 39.8Improved
add DynConv without parallel supervision (ablation)vanilla query-based mask head · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 37.4 -> 38.8Improved
add parallel supervision to vanilla mask head (ablation)vanilla query-based mask head · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 37.4 -> 37.9Improved
add HTC mask flow to vanilla query-based mask head (ablation)query-based cascade mask head · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 37.9 -> 38.9 mask AP; FPS 11.1 -> 6.0Mixed
add HTC mask flow on top of DynConv (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 39.8 -> 40.0 mask AP; FPS 10.5 -> 5.4No clear effect
share query and MSA between detection and segmentation (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 38.1 -> 39.8Improved
share query and MSA between detection and segmentation (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)box AP: 43.4 -> 44.5Improved
share only MSA (not query) (ablation)QueryInst · COCO val2017 (R50-FPN, 100 queries, mask head ablation)mask AP: 38.1 -> 38.3No clear effect
use QueryInst instead of HTCHTC · COCO test-dev (R101-FPN, 36 ep, 480-800 w/ crop)mask AP: 40.8 -> 42.8Improved
increase queries from 100 to 300QueryInst · COCO test-dev (R50-FPN, 36 ep, 640-800)mask AP: 40.1 -> 40.6; FPS 10.5 -> 7.0Improved
use QueryInst on Sparse R-CNN detector for box predictionSparse R-CNN · COCO test-dev (R101-FPN, 36 ep, 480-800 w/ crop)box AP: 46.3 -> 48.1Improved

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