LoGeR – 3D reconstruction from extremely long videos (DeepMind, UC Berkeley)
LoGeR enables dense 3D reconstruction of extremely long videos by processing them in chunks with a hybrid memory module, addressing quadratic complexity issues. It utilizes Sliding Window Attention for local alignment and Test-Time Training to maintain global consistency, effectively reducing drift over sequences of up to 19,000 frames.
loger-project.github.io
4 min
3/10/2026
LoGeR – 3D reconstruction from extremely long videos (DeepMind, UC Berkeley)
LoGeR enables dense 3D reconstruction of extremely long videos by processing them in chunks with a hybrid memory module, addressing quadratic complexity issues. It utilizes Sliding Window Attention for local alignment and Test-Time Training to maintain global consistency, effectively reducing drift over sequences of up to 19,000 frames.
loger-project.github.io
4 min
3/10/2026
LoGeR – 3D reconstruction from extremely long videos (DeepMind, UC Berkeley)
LoGeR enables dense 3D reconstruction of extremely long videos by processing them in chunks with a hybrid memory module, addressing quadratic complexity issues. It utilizes Sliding Window Attention for local alignment and Test-Time Training to maintain global consistency, effectively reducing drift over sequences of up to 19,000 frames.
loger-project.github.io
4 min
3/10/2026
No more articles to load