AI Paper Atlas

Research map / Video / Video generation

SAVA-X: Ego-to-Exo Imitation Error Detection via Scene-Adaptive View Alignment and Bidirectional Cross View Fusion

Xiang Li, Heqian Qiu, Lanxiao Wang, Benliu Qiu, Fanman Meng, Linfeng Xu, Hongliang Li

CVPR 2026 · Video › Video generation · approach: Trajectory-guided motion control · 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
use exo demonstration plus ego video instead of ego-only input (ablation)PDVC · EgoMe valMean AUPRC: 12.79 -> 18.88Improved
add adaptive sampling (AS) to the backbone (ablation)PDVC backbone · EgoMe valMean AUPRC: 18.88 -> 20.90Improved
add scene-adaptive view embeddings (SVE) (ablation)PDVC backbone · EgoMe valMean AUPRC: 18.88 -> 21.29Improved
add bidirectional cross-attention fusion (BiX) (ablation)PDVC backbone · EgoMe valMean AUPRC: 18.88 -> 21.06Improved
add all three modules together (ablation)PDVC backbone · EgoMe valMean AUPRC: 18.88 -> 22.36Improved
add AS to SVE+BiX (full model vs SVE+BiX) (ablation)SAVA-X · EgoMe valMean AUPRC: 22.33 -> 22.36No clear effect
add adaptive sampling across input frame rates (ablation)channel-concat baseline · EgoMe valAUPRC@0.5: 20.61 -> 22.39Improved
add Lsel and Lvic regularizers to adaptive sampling (ablation)adaptive sampling · EgoMe valAUPRC@0.5: 22.39 -> 22.85Improved
adaptive sampling at 20 fps input (ablation)channel-concat baseline · EgoMe val 20 fpsAUPRC@0.5: 21.09 -> 21.45Mixed
replace temporal concatenation fusion with bidirectional cross-attention (ablation)fusion · EgoMe valMean AUPRC: 19.27 -> 21.06Improved
use bidirectional instead of Exo->Ego only cross-attention (ablation)BiX · EgoMe valMean AUPRC: 20.73 -> 21.06Improved
use bidirectional instead of Ego->Exo only cross-attention (ablation)BiX · EgoMe valMean AUPRC: 19.48 -> 21.06Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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