AI Paper Atlas

Research map / Detection and segmentation / CNN features

From Pixels to Objects: A Hierarchical Approach for Part and Object Segmentation Using Local and Global Aggregation

Yunfei Xie, Cihang Xie, Alan Yuille, Jieru Mei

ECCV 2024 · Detection and segmentation › CNN features · approach: Boundary-aware instance segmentation · 0 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 bilinear upsampling with association-aware upsampling (ablation)LGFormer · PartImageNet valobject mIoU: 73.2 -> 79.8Improved
replace bilinear upsampling with association-aware upsampling (ablation)LGFormer · PartImageNet valpart mIoU: 65.3 -> 67.4Improved
use 64 group tokens instead of 256 (ablation)LGFormer · PartImageNet valobject mIoU: 78.2 -> 79.8Improved
use 64 group tokens instead of 16 (ablation)LGFormer · PartImageNet valobject mIoU: 78.6 -> 79.8Improved
initialize group tokens by average pooling instead of learnable tokens (ablation)LGFormer · PartImageNet valobject mIoU: 79.3 -> 79.8Improved
use 3 GCA stages instead of 2 (ablation)LGFormer · PartImageNet valobject mIoU: 77.7 -> 79.8Improved
use 3 GCA stages instead of 4 (ablation)LGFormer · PartImageNet valobject mIoU: 78.5 -> 79.8Improved
train part and object segmentation separately instead of jointlyLGFormer · PartImageNet valpart mIoU: 67.4 -> 69.4Improved
evaluate on 20-40% occluded imagesLGFormer · Occluded-PartImageNet-v1part/object mIoU: part -8.0, object -15.5Hurt
LGFormer vs Compositor under joint trainingPascal-Part · Pascal-Part valobject mIoU: 83.1 -> 85.1Improved

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