Mamba3R: Fixed-Capacity Selective State Memory for Long-Sequence 3D Reconstruction
cs.CV
Submitted: 2026-11-03
Updated: 2026-11-13
Terminology
Sources
- Streaming 4D Visual Geometry Transformer
- TTT3R: 3D Reconstruction as Test-Time Training
- Point3R: Streaming 3D Reconstruction with Explicit Spatial Pointer Memory
- ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training
- TTSA3R: Training-Free Temporal-Spatial Adaptive Persistent State for Streaming 3D Reconstruction
- Mem3R: Streaming 3D Reconstruction with Hybrid Memory
- LoGeR: Long-Context Geometric Reconstruction with Hybrid Memory
- Mamba-VGGT: Persistent Long-Sequence Video Geometry Grounded Transformer via External Sliding Window Mamba Memory
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs