AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Foreground-Aware Token Routing Vision Transformer for Real-Time Satellite Video Tracking

Jiahao Wang, Fang Liu, Licheng Jiao, Shuo Li, Hao Wang, Lingling Li, Xinyi Wang, Xu Liu

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Hybrid and interpretable ViT designs

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 ASM instead of linear mapping (ablation)FATrack-Base · SV248SSuccess: 0.432 -> 0.468Improved
add ASM instead of linear mapping (ablation)FATrack-Base · VISOSR: 0.245 -> 0.286Improved
use ASM instead of bilinear interpolation (ablation)FATrack-Base · SV248SSuccess: 0.421 -> 0.468Improved
use ASM instead of zero padding (ablation)FATrack-Base · SatSOTSuccess: 0.425 -> 0.470Improved
add Shrinkage-Expansion token routing (ablation)FATrack-Base · VISOSR: 0.167 -> 0.286Improved
add Shrinkage-Expansion token routing (ablation)FATrack-Base · SV248SFPS: 100 -> 142Improved
apply shrink-expand only at first and last layers (ablation)FATrack-Base · SV248SSuccess: 0.468 -> 0.433Hurt
use centrality-guided filtering instead of attention-guided (ablation)OSTrack-style tracker · SV248SSR: 39.9 -> 46.8Improved
use centrality-guided filtering instead of no filtering (ablation)FATrack-Base · SV248SMACs (G): 29.0 -> 17.3Improved
use random token filtering instead of centrality-guided (ablation)FATrack-Base · SV248SSR: 46.8 -> 38.1Hurt

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