AI Paper Atlas

Research map / Video / Video generation

LoGeR: Long-Context Geometric Reconstruction with Hybrid Memory

Junyi Zhang, Charles Herrmann, Junhwa Hur, Chen Sun, Ming-Hsuan Yang, Forrester Cole, Trevor Darrell, Deqing Sun

ECCV 2026 · Video › Video generation · approach: Autoregressive long video generation · 45 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
remove TTT layers from the hybrid memory (ablation)LoGeR · ScanNet subset 1000fATE: 0.162 -> 0.107Improved
remove SWA layers (ablation)LoGeR · ScanNet subset 1000fATE: 0.143 -> 0.107Improved
add TTT on top of SWA (ablation)LoGeR · TUM 1000fATE: 0.079 -> 0.050Improved
add SWA on top of TTT (ablation)LoGeR · TUM 1000fATE: 0.053 -> 0.050Improved
add five large-scale datasets to the training mixture (ablation)LoGeR · ScanNet subset 1000fATE: 0.156 -> 0.107Improved
add progressive curriculum training (ablation)LoGeR · ScanNet subset 1000fATE: 0.133 -> 0.107Improved
add progressive curriculum training (ablation)LoGeR* · TUM 500fATE: 0.029 -> 0.031Hurt
add feedforward SE(3) pose alignment (LoGeR* vs LoGeR)LoGeR · KITTI avgATE (m): 25.44 -> 18.65Improved
add hybrid memory over chunk stitching (Pi3-Chunk -> LoGeR)Pi3-Chunk · KITTI avgATE (m): 52.07 -> 25.44Improved
add hybrid memory over chunk stitchingPi3-Chunk · VBR avgATE (m): 70.09 -> 36.16Improved
chunk stitching vs LoGeR on short horizonsLoGeR · VBR ~1k framesATENo clear effect
smaller inference chunk size (64 -> 32)LoGeR · 500-frame efficiencyspeed / memory: 9.3 -> 12.1 FPS; 27.2 -> 18.1 GBMixed

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