AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

SARA: Structure-Aware Riemannian-Guided Alignment for Drone Image-Text Retrieval

Keyu Lu, Qing Ma, Zhenyu Lu, Cong Bai

ECCV 2026 · Vision-language and multimodal › Cross-modal learning · approach: Probabilistic embeddings for cross-modal retrieval · 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
add visual structural features (VSF) from SPD covariance (ablation)semantic baseline · ERA DITmR: 44.11 -> 45.12Improved
add visual structural features (VSF) from SPD covariance (ablation)semantic baseline · UDV DITmR: 21.02 -> 21.35Improved
add textual structural features (TSF) alone (ablation)semantic baseline · ERA DITmR: 44.11 -> 36.86Hurt
add textual structural features (TSF) alone (ablation)semantic baseline · UDV DITmR: 21.02 -> 20.26Hurt
add TSF on top of VSF (ablation)MSFE · ERA DITmR: 45.12 -> 45.43Improved
add marginal distribution alignment (MDA) (ablation)MSFE (VSF+TSF) · ERA DITmR: 45.43 -> 46.62Improved
add marginal distribution alignment (MDA) (ablation)MSFE (VSF+TSF) · UDV DITmR: 21.57 -> 21.79Improved
add conditional distribution alignment (CDA) (ablation)MSFE (VSF+TSF) · ERA DITmR: 45.43 -> 46.81Improved
add CDA on top of MDA (full LESA) (ablation)SARA · ERA DITmR: 46.62 -> 48.36Improved
add MDA and CDA without structural features (ablation)semantic baseline · ERA DITmR: 44.11 -> 45.50Improved
set structural loss weight alpha to 0.7 instead of 1.0 (ablation)SARA · ERA DITmR: 43.52 -> 48.36Improved
set structural loss weight alpha to 0.7 instead of 0.01 (ablation)SARA · ERA DITmR: 45.23 -> 48.36Improved

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