AI Paper Atlas

Research map / Detection and segmentation / CNN features

GMNet: Graph Matching Network for Large Scale Part Semantic Segmentation in the Wild

Umberto Michieli, Edoardo Borsato, Luca Rossi, Pietro Zanuttigh

ECCV 2020 · Detection and segmentation › CNN features · approach: Joint detection and instance segmentation · 40 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 reconstruction loss on object-level segmentation maps (ablation)Deeplab-v3 baseline · Pascal-Part-58mIoU: 54.4 -> 55.2Improved
add semantic embedding network S (ablation)GMNet · Pascal-Part-58mIoU: 55.2 -> 58.4Improved
add weighted graph matching loss (ablation)GMNet · Pascal-Part-58mIoU: 58.4 -> 59.0Improved
add unweighted binary graph matching (ablation)GMNet · Pascal-Part-58mIoU: 58.4 -> 58.7Improved
use weighted instead of unweighted graph (ablation)GMNet · Pascal-Part-58mIoU: 58.7 -> 59.0Improved
concatenate S features at multiple resolutions instead of a single one (ablation)GMNet · Pascal-Part-58mIoU: 58.7 -> 59.0Improved
use embedding network S instead of directly concatenating object maps (ablation)GMNet · Pascal-Part-58mIoU: 55.7 -> 59.0Improved
use S instead of conditioning decoder with object encoder features (ablation)GMNet · Pascal-Part-58mIoU: 55.7 -> 59.0Improved
feed ground-truth object maps to S (upper bound) (ablation)GMNet · Pascal-Part-58mIoU: 59.0 -> 65.6Improved
full GMNet versus Deeplab-v3 baselineGMNet · Pascal-Part-108mIoU: 41.3 -> 45.8Improved
full GMNet versus BSANetGMNet · Pascal-Part-108mIoU: 42.9 -> 45.8Improved
full GMNet versus BSANetGMNet · Pascal-Part-58mIoU: 58.2 -> 59.0Improved

Later papers that compared with this method

3 later accepted papers report GMNet as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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