AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Enhancing Mamba Decoder with Bidirectional Interaction in Multi-Task Dense Prediction

Mang Cao, Sanping Zhou, Yizhe Li, Ye Deng, Wenli Huang, Le Wang

ICCV 2025 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 1 citations (OpenAlex)

Read the paper

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 BI-Scan to MTMamba baseline (ablation)BIM · NYUD-v2MTL gain Delta_m: +2.38 -> +4.02Improved
add MS-Scan to MTMamba baseline (ablation)BIM · NYUD-v2MTL gain Delta_m: +2.38 -> +4.00Improved
combine BI-Scan and MS-Scan (ablation)BIM · NYUD-v2MTL gain Delta_m: +4.02 -> +4.83Improved
use both task-first and position-first scan modes (ablation)BI-Scan · NYUD-v2Semseg mIoU: 56.36 / 56.96 -> 57.11Improved
order TF before PF rather than PF before TF (ablation)BI-Scan · NYUD-v2Semseg mIoU: 56.55 -> 57.11Improved
use bidirectional instead of unidirectional cross-task scan (ablation)BI-Scan · NYUD-v2overall performanceImproved
random task order instead of fixed order (ablation)BI-Scan · NYUD-v2MTL gain Delta_m: +4.02 -> +3.17Hurt
vary fixed task order (ablation)BI-Scan · NYUD-v2MTL gain Delta_m: +3.72 to +4.20No clear effect
MS-Scan scales {1,4} versus {1,2} or {1,3} (ablation)MS-Scan · NYUD-v2MTL gain Delta_m: +4.68 / +4.62 -> +4.83Improved
use more scales across MFRs (Type 3) instead of {1,4} (ablation)MS-Scan · NYUD-v2MTL gain Delta_m: +4.83 -> +4.35Hurt
dilated sampling in MS-Scan (DBIM)BIM · NYUD-v2FLOPs (G) / Semseg mIoU: 547 -> 495 FLOPs; 57.40 -> 56.41 mIoUMixed
replace MTMamba decoder interaction with BIMSwin-Base encoder · NYUD-v2Semseg mIoU: 53.62 -> 54.32Improved

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