AI Paper Atlas

Research map / Detection and segmentation / CNN features

Structure-Aware Representation Distillation for Tiny-Dense Object Segmentation

Xuesong Liu, Anke Xu, Wenbo Cao, Emmett Ientilucci

CVPR 2026 · Detection and segmentation › CNN features · approach: Context modeling for semantic segmentation

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 logit-level KD with uniform feature representation distillation (ablation)ResNet-50 student, Swin-L teacher · RockFragmIoU: 52.3 -> 53.8Improved
add structure-aware weighting (E+C+D) to uniform SARD (ablation)SARD ResNet-50 student · RockFragmIoU: 53.8 -> 58.6Improved
add structure-aware weighting (E+C+D) to uniform SARD (ablation)SARD ResNet-50 student · RockFragbIoU: 38.5 -> 45.2Improved
use boundary-only weighting instead of full structure score (ablation)SARD · RockFragmIoU: 58.6 -> 56.2Hurt
use density-only weighting instead of full structure score (ablation)SARD · RockFragmIoU: 58.6 -> 55.4Hurt
add density term D to E+C weighting (ablation)SARD · RockFragmIoU: 57.8 -> 58.6Improved
use SARD instead of CWD, Swin-L to Swin-TSwin-T student · RockFragmIoU: 55.9 -> 60.2Improved
use SARD instead of CWD, Swin-L to Swin-TSwin-T student · CityscapesmIoU: 78.9 -> 80.3Improved
use SARD instead of CWD, Swin-B to Swin-SSwin-S student · CityscapesmIoU: 79.4 -> 80.1Improved
use SARD instead of CWD, Swin-B to Swin-SSwin-S student · RockFragbIoU: 42.8 -> 46.5Improved
use SARD instead of CWD across heterogeneous teacher-student pairsViT-T, EfficientSAM-Ti, M2F-R50 students · Cityscapes/ADE20K/RockFrag averagemIoU: +1.5 to +1.8 mIoU, +2.3 to +2.7 bIoUImproved
distill ResNet-50 with SARD instead of CWDResNet-50 student · RockFragmIoU: 54.1 -> 58.6Improved

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