A Straight Line Is Not a Camera Pose: Auditing Short-Clip Pose Benchmarks
cs.CV, cs.RO
Submitted: 2026-10-31
Updated: 2026-11-13
Comments: 39 pages, 9 figures
Code: https://github.com/MichaelGrupp/evo
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Noctif3R: Feed-Forward Monocular Real-Time SLAM for Photon-Limited Scenes on Embedded Hardware
- E3D-Bench: A Benchmark for End-to-End 3D Geometric Foundation Models
- Comparing Trajectories from Positions Alone: Curvature-Based Time Alignment and Drift Error Metric
- The Benchmark Lottery
- VLRC: Vision-Language Reprojection Consistency as a scalable signal for better feed-forward 3D pretraining
- SS3D: End2End Self-Supervised 3D from Web Videos
- MONA: Moving Object Detection from Videos Shot by Dynamic Camera
- G-CUT3R: Guided 3D Reconstruction with Camera and Depth Prior Integration
- Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels
- UAVFF3D: A Geometry-Aware Benchmark for Feed-Forward UAV 3D Reconstruction
- Video Generation Models are General-Purpose Vision Learners
- PURE: Refining Feed-forward 3D Reconstruction with Unlabeled Post-training
- SpatialTrackerV2: 3D Point Tracking Made Easy
- LASER: Layer-wise Scale Alignment for Training-Free Streaming 4D Reconstruction
- ARROW: Arbitrary Reconstruction and Tracking of 4D Observations in the Wild
- Efficiently Reconstructing Dynamic Scenes One D4RT at a Time
- HD-VGGT: High-Resolution Visual Geometry Transformer
- Less Context, Better Geometry: Masked Geometric Encoder for Robust 3D Foundation Models
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